CN115933504B - Travel control system, travel control method and apparatus - Google Patents

Abstract

The embodiment of the application provides a running control system, a running control method and a running control device, wherein the running control system comprises: the first processor is connected with each second processor respectively, wherein: the first processor is used for determining a target algorithm corresponding to the target running control scene under the condition that running of the running equipment is required to be controlled according to the target running control scene, and screening a main processor and a slave processor which are used for running the target algorithm from the plurality of second processors, wherein the main processor is used for running the target algorithm and outputting a target operation result; the first processor is also used for acquiring a target operation result and generating a running control instruction for the running equipment according to the target operation result. Through the method and the device, the problem that the control efficiency of the running control system to the running equipment is low is solved, and the effect of improving the control efficiency of the running control system to the running equipment is achieved.

Description

Driving control system, driving control method and device

technical field

The embodiments of the present application relate to the field of data processing, and in particular, relate to a driving control system, a driving control method and a device.

Background technique

Autonomous driving, that is, the use of sensors and computer systems to realize unmanned intelligent driving, has a history of several decades in the 20th century, and it has shown a trend of being close to practicality in the early 21st century. With the continuous development of computer, pattern recognition, artificial intelligence and other technologies, more and more computer control technologies are applied to automobiles, forming automatic driving technology, which has become a hot spot of public concern. According to the SAE J3016 standard formulated by the Society of Automotive Engineers International in 2014, automotive automation systems can be divided into L0 (no automation), L1 (assisted driving), L2 (partial automation), L3 (conditional automation), L4 (high automation) and L 5 (full automation) six levels. This standard has not only been adopted by the U.S. Department of Transportation as a federal standard, but has also become a common standard for the global automotive industry to assess the level of autonomous driving. At present, the automatic driving system referred to in the industry is mainly the L4 automotive automation system. Its driving decision-making is mainly controlled by the car itself, not by humans. This requires higher precision, more reliable perception, decision-making planning, and control in the on-board computing system. Algorithm, which requires 100-1000 TOPS (processor computing power unit) high computing power platform support. At present, the autonomous driving algorithms of many automakers still continue the ecological development of NVIDIA technology, but NVIDIA’s high-computing chips that meet automotive regulations have not yet been mass-produced, and the vehicle-mounted computing system mainly uses CPU (Central Processing Unit, central processing unit) + GPU (Graphics Processing Unit) , graphics processor) industrial computer and NVIDIA Jetson Xavier embedded computing module, the CPU+GPU industrial computer computing system can provide high computing power, but it cannot meet the reliability requirements of the vehicle. In order to improve the safety and reliability of the system, generally The method is to use multiple industrial computers for fault-tolerant calculation, that is, multiple industrial computers are mutually redundant. When one industrial computer fails, other redundant industrial computers can take over the work of the failed industrial computer. Although the method meets the reliability requirements, it is followed by high power consumption, and the method of multiple industrial computers is not conducive to the vehicle.

Contents of the invention

Embodiments of the present application provide a driving control system, a driving control method and a device, so as to at least solve the problem in the related art that the driving control system controls driving equipment with low efficiency.

According to an embodiment of the present application, a driving control system is provided, including: a first processor and a plurality of second processors, wherein the first processor is respectively connected to each second processor, wherein: the first a processor, configured to determine the target algorithm corresponding to the target driving control scenario when the driving device needs to be controlled according to the target driving control scenario, and select the main processing for running the target algorithm from multiple second processors The processor and the slave processor, wherein the target algorithm is used to determine the driving environment of the driving device; the main processor is used to run the target algorithm and output the target calculation result, wherein the target calculation result is used to characterize the driving environment of the driving device; the slave processing The first processor is used to take over the main processor to run the target algorithm and output the target calculation result corresponding to the target algorithm when the main processor is in a fault state; the first processor is also used to obtain the target calculation result and As a result, a driving control instruction for the driving equipment is generated, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment.

In an exemplary embodiment, the travel control system further includes a third processor, wherein the third processor is respectively connected to the first processor and each second processor, and the third processor is used to monitor the first processing The first operating state of the device and the second operating state of the second processor, and determine the driving automation level of the driving device according to the first operating state and the second operating state.

In an exemplary embodiment, the first processor and each second processor are connected through multiple data transmission links, and the first processor is used to monitor the first data transmission in the multiple data transmission links The link state of the link, and in the case that the first data transmission link is in a fault state, determine the second data transmission link except the first data transmission link from the multiple data transmission links, and pass Data is transmitted between the second data transmission link and the second processor, wherein the first data transmission link is a link currently transmitting data between the first processor and the second processor.

In an exemplary embodiment, the driving control system further includes an environment sensing device, wherein the sensing device is respectively connected to each second processor, and the sensing device is used to collect environmental information of the driving environment and synchronously transmit the environmental information to master and slave processors.

According to an embodiment of the present application, a driving control method is provided, which is applied to the first processor in the driving control system, including: when the driving equipment needs to be controlled according to the target driving control scene, determine and target driving control A target algorithm corresponding to the scene, wherein the target algorithm is used to determine the driving environment of the driving device; a master processor and a slave processor are selected from multiple second processors deployed in the driving control system, wherein the first processors are respectively Connected with each second processor, the main processor is used to run the target algorithm and output the target calculation result, the target calculation result is used to characterize the driving environment of the driving equipment, and the slave processor is used to take over when the main processor is in a fault state The main processor runs the target algorithm and outputs the target calculation result corresponding to the target algorithm; obtains the target calculation result; generates a driving control command for the driving device according to the target calculation result, wherein the driving control command is used to instruct the driving device in the driving environment driving status.

In an exemplary embodiment, selecting a master processor and a slave processor for running the target algorithm from a plurality of second processors includes: determining a target computing power value required for running the target algorithm; The value filters the primary and secondary processors from multiple secondary processors.

In an exemplary embodiment, selecting the master processor and the slave processor from multiple second processors according to the target computing power value includes: when the target computing power value is less than or equal to the preset computing power value, Select the target processor from multiple second processors as the main processor, wherein the target processor is a processor whose computing power value is greater than or equal to the target computing power value among the multiple second processors, and the preset computing power value is half of the sum of the computing power values of the multiple second processors; among the multiple second processors, the processors other than the main processor are determined as slave processors.

In an exemplary embodiment, the selection of the master processor and the slave processors from the plurality of second processors according to the target computing power value includes: when the target computing power value is greater than a preset computing power value, determining how many The algorithm importance of each target algorithm in each target algorithm, where the algorithm importance is used to indicate the impact of the target calculation results corresponding to the target algorithm on the accuracy of the driving control instruction; according to the algorithm importance of each target algorithm from many Candidate algorithms are determined from target algorithms; according to the computing power values and candidate algorithms of multiple second processors, the main processor and the secondary processor are selected from the second processors, wherein the main processor and the secondary processor are selected from the second processor. The computing power values are all greater than or equal to the computing power values required to run the candidate algorithm.

In an exemplary embodiment, determining a candidate algorithm from multiple target algorithms according to the algorithm importance of each target algorithm includes: selecting an algorithm whose algorithm importance is greater than the target importance from the multiple target algorithms as a candidate algorithm , wherein the target importance is determined according to the algorithmic importance of multiple target algorithms.

In an exemplary embodiment, selecting the master processor and the slave processor from the second processors according to the computing power values and candidate algorithms of the multiple second processors includes: screening the multiple second processors The processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm is used as the main processor, and is selected from the processors other than the main processor included in the multiple second processors with a computing power value greater than or equal to the candidate algorithm. The processor of the computing power value of the algorithm acts as a slave processor.

In an exemplary embodiment, before obtaining the target operation result, the method further includes: adding the first algorithm startup information and the second algorithm startup information to the startup file, wherein the first algorithm startup information is the The second algorithm startup information is the information of the algorithm that needs to be run on the secondary processor, and the startup file is used to run the algorithm indicated by the algorithm startup information on the processor indicated by the algorithm startup information.

In an exemplary embodiment, before obtaining the target operation result, the method further includes: determining a reference data subscription relationship between any two target algorithms, wherein the reference data subscription relationship is used to indicate that multiple target algorithms are running , the target algorithm running later is used to perform data operations based on the output of the target algorithm running earlier; according to the reference data subscription relationship, modify the target data between the target algorithms running on the master processor and the slave processor respectively Subscription relationship.

In an exemplary embodiment, after modifying the data subscription relationship between the master processor and the slave processor for running each target algorithm according to the data subscription relationship, the method further includes: when the target data subscription relationship indicates that the first algorithm subscribes The operation result of the second algorithm is obtained, and the second algorithm is configured to call the communication monitoring algorithm to monitor the target communication state of the second algorithm running on the main processor under the condition that the main processor and the slave processor are running, wherein, The target algorithm includes a first algorithm and a second algorithm, and the target communication state includes a communication frequency state and a communication time state of the second algorithm; according to the target communication state, the first operation result and the second operation result are selected for transmission to the second algorithm. An operation result of an algorithm, wherein the first operation result is the result output by the master processor running the second algorithm, and the second operation result is the result output by the slave processor running the second algorithm.

In an exemplary embodiment, according to the target communication state, the operation result for transmission to the first algorithm is selected from the first operation result and the second operation result, including: the target communication state indicates that the main processor runs on the When the second algorithm is in an abnormal communication state, determining that the second operation result is an operation result for transmission to the first algorithm; when the target communication state indicates that the second algorithm running on the main processor is in a normal communication state, Invoke the error monitoring algorithm to monitor the error value of the first calculation result; in the case that the error value of the first calculation result is less than or equal to the target error value, determine that the first calculation result is the calculation result for transmission to the first algorithm; When the error value of an operation result is greater than the target error value, the second operation result is determined to be the operation result to be transmitted to the first algorithm.

According to another embodiment of the present application, there is provided a driving control method applied to a third processor in a driving control system, comprising: monitoring the first operating state of the first processor deployed in the driving control system and a plurality of The second operating state of the second processor, wherein the first processor is respectively connected to each second processor, the third processor is connected to the first processor and each second processor, and the first processor is used for When it is necessary to control the running of the driving device according to the target driving control scene, determine the target algorithm corresponding to the target driving control scene, and select the main processor and the slave processor for running the target algorithm from a plurality of second processors, The target algorithm is used to determine the driving environment of the driving equipment; the main processor is used to run the target algorithm and output the target calculation results, and the target calculation results are used to characterize the driving environment of the driving equipment; the slave processor is used to , take over from the main processor to run the target algorithm, and output the target calculation result corresponding to the target algorithm; the first processor is also used to obtain the target calculation result, and generate a driving control command for the driving device according to the target calculation result, and the driving control command is used for Indicate the driving state of the driving device in the driving environment; determine the driving automation level of the driving device according to the first running state and the second running state.

In an exemplary embodiment, determining the driving automation level of the driving device according to the first operating state and the second operating state includes: determining the master processor and the slave processor through the first method, and satisfying the first predetermined In the case of setting conditions, it is determined that the driving automation level of the driving equipment is the assisted driving level, wherein the first preset condition includes one of the following: the first operating state indicates that the first processor is in a fault state, and the second operating state indicates that a plurality of The number of processors in the fault state in the second processor is greater than or equal to half of the total number of multiple second processors, wherein the first method includes: the target computing power value required to run the target algorithm is less than or equal to the preset computing power value, select the target processor from multiple second processors as the master processor, determine the processors other than the master processor among the multiple second processors as slave processors, and target processors as multiple For a processor whose computing power value is greater than or equal to the target computing power value among the second processors, the preset computing power value is half of the sum of the computing power of multiple second processors; Determined in the first way, and in the case of meeting the second preset condition, it is determined that the driving automation level of the driving equipment is a highly automated driving level, wherein the second preset condition includes one of the following: the first operating state indicates the first The processors are in a normal state; the second running state indicates that the number of processors in a failure state among the plurality of second processors is less than half of the total number of the plurality of second processors.

In an exemplary embodiment, after determining the driving automation level of the traveling device according to the first operating state and the second operating state, the method further includes: after the master processor and the slave processor are determined through the second method, and at When the third preset condition is met, it is determined that the driving automation level of the driving equipment is an assisted driving level, wherein the third preset condition includes one of the following: the first operating state is used to indicate that the first processor is in an abnormal state, and the second The second running state is used to indicate that the master processor and the slave processor have failures at the same time, wherein the second way is to determine the number of targets among multiple target algorithms when the target computing power value required to run the target algorithm is greater than the preset computing power value. Algorithm importance of each target algorithm, where the algorithm importance is used to indicate the impact of the target calculation results corresponding to the target algorithm on the accuracy of the driving control command, according to the algorithm importance of each target algorithm from multiple target algorithms Determine the candidate algorithm, and select the main processor and the slave processor from the second processors according to the computing power values of multiple second processors and the candidate algorithm, and the computing power of the main processor and the slave processor is greater than or equal to The computing power value of the candidate algorithm; the main processor and the slave processor are determined by the second method, and in the case of meeting the fourth preset condition, determine that the driving automation level of the driving equipment is a partially automated driving level, wherein, The fourth preset condition includes: the first operating state is used to indicate that the first processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processor are not in a fault state at the same time, the third processor is in a fault state, The third processor is a processor other than the main processor and the slave processor among the plurality of second processors; the master processor and the slave processor are determined by the second method, and the fifth preset condition is met In this case, it is determined that the driving automation level of the driving equipment is a highly automated driving level, wherein the fifth preset condition includes: the first operating state is used to indicate that the first processor is in a normal state, and the second operating state is used to indicate that the main processor Unlike the slave processors which are simultaneously in a fault state, the third processor is in a normal state.

In an exemplary embodiment, after determining the driving automation level of the traveling device according to the first operating state and the second operating state, the method further includes: when the driving automation level is lower than the target level, sending a control instruction to the traveling device , the control command is used to instruct the driving equipment to pull over and stop.

According to another embodiment of the present application, a driving control device is provided, which is applied to the first processor in the driving control system, including: a first determination module, used to control the driving of the driving equipment according to the target driving control scene In this case, determine the target algorithm corresponding to the target driving control scenario, wherein the target algorithm is used to determine the driving environment of the driving device; the first screening module is used to filter out the multiple second processors deployed in the driving control system a master processor and a slave processor, wherein the first processor is respectively connected to each second processor, the master processor is used to run the target algorithm, and outputs the target calculation result, and the target calculation result is used to characterize the driving environment of the driving device, and the slave The processor is used to replace the main processor to run the target algorithm and output the target calculation result corresponding to the target algorithm when the main processor is in a fault state; the acquisition module is used to obtain the target calculation result; the generation module is used to perform the target calculation according to the target algorithm As a result, a driving control instruction for the driving equipment is generated, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment.

According to another embodiment of the present application, a driving control device is provided, which is applied to the third processor in the driving control system, including: a monitoring module, used to monitor the first processor of the first processor deployed in the driving control system An operating state and a second operating state of a plurality of second processors, wherein the first processor is respectively connected to each second processor, the third processor is connected to the first processor and each second processor, and the first processor is connected to each second processor. A processor is used to determine the target algorithm corresponding to the target driving control scene when the driving device needs to be controlled according to the target driving control scene, and select a main processor for running the target algorithm from multiple second processors And from the processor, the target algorithm is used to determine the driving environment of the driving device; the main processor is used to run the target algorithm, output the target calculation result, and the target calculation result is used to characterize the driving environment of the driving device; In the case of a fault state, take over the main processor to run the target algorithm, and output the target calculation result corresponding to the target algorithm; the first processor is also used to obtain the target calculation result, and generate a driving control command for the driving equipment according to the target calculation result, The driving control instruction is used to indicate the driving state of the driving device in the driving environment; the third determination module is used to determine the driving automation level of the driving device according to the first running state and the second running state.

According to yet another embodiment of the present application, a computer-readable storage medium is also provided, and a computer program is stored in the computer-readable storage medium, wherein the computer program is set to execute any one of the above-mentioned method embodiments when running. step.

According to yet another embodiment of the present application, there is also provided an electronic device, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to perform the steps in any one of the above method embodiments .

According to the present application, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each of the second processors, and when it is necessary to control the driving of the driving equipment according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

Description of drawings

FIG. 1 is a structural diagram of an optional driving control system according to an embodiment of the present application;

FIG. 2 is an architecture diagram of an optional driving control system according to an embodiment of the present application;

Fig. 3 is a fault-tolerant schematic diagram of an optional travel control system according to an embodiment of the present application;

Fig. 4 is a block diagram of the hardware structure of a mobile terminal of a driving control method according to an embodiment of the present application;

FIG. 5 is a first flowchart of a driving control method according to an embodiment of the present application;

Fig. 6 is an optional driving control flow chart according to an embodiment of the present application;

FIG. 7 is a second flow chart of a driving control method according to an embodiment of the present application;

Fig. 8 is a structural block diagram 1 of a driving control device according to an embodiment of the present application;

FIG. 9 is a second structural block diagram of a driving control device according to an embodiment of the present application.

Detailed ways

Embodiments of the present application will be described in detail below with reference to the drawings and in combination with the embodiments.

It should be noted that the terms "first" and "second" in the description and claims of the present application and the above drawings are used to distinguish similar objects, but not necessarily used to describe a specific sequence or sequence.

Fig. 1 is a structural diagram of an optional driving control system according to an embodiment of the present application. As shown in Fig. 1 , the driving control system includes: a first processor 12 and a plurality of second processors 14, wherein the first The processor 12 is respectively connected to each second processor 14, wherein:

The first processor 12 is used to determine the target algorithm corresponding to the target driving control scenario when it is necessary to control the driving of the driving device according to the target driving control scenario, and to filter out the operating target from multiple second processors 14. a master processor and a slave processor of an algorithm, wherein the target algorithm is used to determine the driving environment of the driving device;

The main processor is used to run the target algorithm and output the target calculation result, wherein the target calculation result is used to represent the driving environment of the driving device;

The slave processor is used to replace the main processor to run the target algorithm when the main processor is in a fault state, and output the target operation result corresponding to the target algorithm;

The first processor 12 is also used to obtain the target calculation result, and generate a driving control instruction for the driving equipment according to the target calculation result, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment.

Based on the above content, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each second processor, and when it is necessary to control the driving of the driving device according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

Optionally, in this embodiment, the selection of the master processor and the slave processor from the second processor may be obtained by filtering the second processor according to the target computing power value required to run the target algorithm, for example, Four second processors with the same computing power are deployed in the driving control system, and the computing power of each second processor is 225 TOPS. When the target computing power required by the target algorithm is lower than 550 TOPS (that is, 2 Orin ( The sum of the computing power of the second processor), the entire autonomous driving algorithm stack can meet the computing power requirements by running on two Orins, and can combine two Orin computing units among the four Orin computing units with the other two Orin computing units One-to-one computing redundancy is formed, that is, two Orins are used as the main processor and the other two Orins are used as slave processors. When any Orin computing device fails, its redundant Orin can take over in real time, so that Orin computing can be fault-tolerant.

Optionally, in this embodiment, when there are multiple target algorithms, selecting the master processor and the slave processor from the second processor may also be based on the target computing power value required to run the target algorithm and the target algorithm The corresponding algorithm importance is obtained by screening the second processor. For example, four second processors with the same computing power are deployed in the driving control system, and the computing power of each second processor is 225 TOPS. When the target algorithm The sum of the target computing power value is larger than 550TOPS (that is, the sum of the computing power of 2 Orins), and the entire algorithm stack needs to run on 3 Orins to meet the computing power requirements. The 4 Orins can be divided into There are 2 parts, of which 3 Orins are used as the main computing unit, and 1 Orin is used as the fault-tolerant computing unit. It forms a one-to-one redundancy with one Orin in the main computing unit, and mainly deals with perception algorithms with high importance of algorithms, such as for For the perception algorithm of the front camera, this fault-tolerant method only implements fault-tolerant processing for key algorithms. These key algorithms all run on one Orin. When the Orin computing device fails, its redundant Orin can take over in real time, so that Orin computing can be fault-tolerant. .

Optionally, in this embodiment, the number of second processors may be set according to the driving control operations performed by the driving control system on the driving equipment. The driving control operations performed by the driving control system are different, and the The number of second processors is different. For example, when the driving control operation performed by the driving control system is lane keeping, the number of second processors is four; when the driving control operation performed by the driving control system is auxiliary braking, the second processor The number of second processors can be 6, or the number of second processors can also be configured as a fixed number, for example, the same number of second processors is configured for driving control systems with different functions, which is not limited in this solution.

Optionally, in this embodiment, the first processor is responsible for logic operations in the driving control process, which may include, but is not limited to, decision planning (generating driving control instructions according to the target calculation results corresponding to the target algorithm), and the second processor Algorithm control, algorithm running status monitoring, second processor running status monitoring and other operations are not limited in this solution.

Optionally, in this embodiment, the target algorithm may include, but is not limited to, an image processing algorithm, a perception algorithm, a perception fusion algorithm, an image recognition algorithm, etc., and each second processor may be configured to run a fixed-function algorithm Or it can also be configured to run an algorithm of any function, for example, an algorithm of a function is pre-stored on each second processor among multiple second processors, so that each second processor can only realize the corresponding function , or each second processor can run any algorithm, and when the second processor needs to run the algorithm, the first processor sends the algorithm to be run to the second processor, which is not limited in this solution .

As an optional embodiment, the travel control system further includes a third processor, wherein the third processor is respectively connected to the first processor and each second processor,

The third processor is configured to monitor the first operating state of the first processor and the second operating state of the second processor, and determine the driving automation level of the driving device according to the first operating state and the second operating state.

Optionally, in this embodiment, the running status may be, but not limited to, used to indicate whether there is a fault in the processor, the data load of the processor, the service blocking status of the processor, etc., which is not limited in this solution.

Optionally, in this embodiment, the driving automation levels include: L0 level (pure manual driving), L1 level (assisted driving), L2 level (partially automated driving), L3 level (conditional automated driving), L4 level (highly autonomous driving), L5 level (fully autonomous driving).

Optionally, in this embodiment, after determining the driving automation level of the driving equipment, the third processor may also send a corresponding control instruction to the chassis control system of the vehicle when the driving automation level is lower than the target automation level, In order to ensure the driving safety of the vehicle, for example, if the target automation level is L2, when the driving automation level of the vehicle is lower than L2, the third processor sends a lift control command to the chassis control system, so that the vehicle pulls over and stops.

As an optional embodiment, the first processor and each second processor are connected through multiple data transmission links,

The first processor is configured to monitor the link state of the first data transmission link among the multiple data transmission links, and when the first data transmission link is in a failure state, select the data transmission link from the multiple data transmission links determining a second data transmission link other than the first data transmission link, and transmitting data between the second data transmission link and the second processor, wherein the first data transmission link is the first processor's current A link to transfer data to and from the second processor.

Optionally, in this embodiment, a fault-tolerant communication link is set between the first processor and each second processor, and a fixed link is used by default between the first processor and the second processor. Data interaction, when one of the links fails, it can automatically switch to other alternative links. For example, the driving control system includes 2 sets of communication networks. The first processor and the second processor can be interconnected through PCIESwitch, using PCIE (one A common bus specification) communication, or through an Ethernet switch, using Ethernet communication, when the PCIE communication physical hardware or communication link fails, it can be switched to the Ethernet communication link, so as to realize the communication between CPU and Orin fault tolerance.

As an optional embodiment, the driving control system further includes an environment sensing device, wherein the sensing device is respectively connected to each second processor,

The perception device is used to collect the environmental information of the driving environment, and transmit the environmental information to the main processor and the slave processor synchronously.

Optionally, in this embodiment, the perception device may include, but is not limited to, a camera device and a radar device (such as millimeter wave radar, lidar, etc.).

Optionally, in this embodiment, the fault tolerance of the sensing data set between the sensing device and the main processor, that is, the same sensing data collected by the same sensing device is transmitted to the main processor at the same time , will also be transmitted to the slave processor of the corresponding setting, for example, to perform fault tolerance on the three main types of sensing data, the first type is camera data, which is directly output to Orin (the second processor); the second type is lidar data , which can also be directly output to Orin (the second processor); the third type is the millimeter-wave radar data, which is read by the MCU (the third processor) through the chassis CAN (Controller Area Network, controller area network bus) port, and then sent To the CPU (the first processor), and then output to Orin (the second processor) through PCIE. In order to ensure the real-time fault tolerance of the system, the data of the same camera, lidar and millimeter wave radar need to be connected to two Orins synchronously to realize fault tolerance. If the system accesses N camera data in total, the first N/2 camera data are sent to both Orin1 (master processor) and Orin3 (slave processor) at the same time. Orin1 and Orin3 can realize real-time synchronization of camera data. If Orin1 fails, Orin3 can take over immediately; the rear N/2 camera data are connected to Orin2 (master processor) and Orin4 (slave processor) at the same time, and Orin2 and Orin4 can realize real-time synchronization of camera data. When Orin2 fails, Orin4 It can take over immediately; and so on, Orin1 and Orin3 read the same lidar and millimeter-wave radar data at the same time, and Orin2 and Orin4 read the same lidar and millimeter-wave radar data at the same time, so that the data input of the three main sensing devices can be realized Synchronization, avoiding the delay overhead caused by switching data reading caused by system failure, thus ensuring the high real-time fault tolerance of the system.

Fig. 2 is an architecture diagram of an optional driving control system according to an embodiment of the present application. As shown in Fig. 2 , the driving control system includes at least sensing devices (cameras, laser radars and millimeter-wave radars), a first processor ( CPU), second processor (Orin) and third processor (MCU), CPU and Orin1, Orin2, Orin3, Orin4 are connected to PCIESwitch, CPU and Orin1, Orin2, Orin3, Orin4 use PCIE3.0 communication, The CPU and Orin1, Orin2, Orin3, and Orin4 are also connected to the Gigabit Ethernet switch at the same time, and the CPU and Orin1, Orin2, Orin3, and Orin4 use Gigabit Ethernet communication, and the CPU computing unit and the MCU (Microcontroller Unit, micro control unit) The safety control unit is interconnected through Gigabit Ethernet, and the MCU is also connected to the Gigabit switch; camera data, which is directly output to Orin, does not need to be transferred through the CPU; the second type is lidar data, which can also be directly output to Orin; The third category is millimeter-wave radar data, which is read by the MCU through the CAN port of the chassis, then sent to the CPU, and then output to Orin through PCIE3.0. Two Orins need to be connected synchronously to achieve fault tolerance.

In the above embodiments, the software and hardware can be implemented through the heterogeneous distributed on-board computing system of the first processor (CPU), the second processor (Orin) and the third processor (MCU) and the automatic driving framework Autoware.Universe Collaborative fault-tolerant design, FIG. 3 is a fault-tolerant schematic diagram of an optional driving control system according to an embodiment of the present application. As shown in FIG. Complete fault tolerance, so as to realize the overall fault tolerance of the system, in which the computing layer, communication layer and data IO layer are fault-tolerant by the hardware system, and the algorithm layer is fault-tolerant by the software system, which at least includes the following contents:

Computing layer fault-tolerant design, this system includes three types of computing chip units, the first type is X86 CPU computing unit, the number is 1; the second type is Orin computing unit (64GB module, computing power is 275TOPS, hereinafter referred to as Orin ), the number is 4, respectively recorded as Orin1, Orin2, Orin3, Orin4; the third type is the MCU unit (hereinafter referred to as MCU), the number is 1. The CPU computing unit mainly processes logical operations, such as autonomous driving decision-making planning, control algorithms, algorithm operation status monitoring algorithms, etc.; Orin1, Orin2, Orin3, and Orin4 mainly process GPU, AI (Artificial Intelligence, artificial intelligence) and other calculations, such as image preprocessing , perception, perception fusion and other algorithms; the MCU unit is mainly used for security control, such as hardware fault monitoring algorithms and communication modules with the chassis. The CPU computing unit is used as the main computing node, configured as the HOST side, and the 4 Orin computing units are used as the slave computing nodes, configured as the Endpoint (hereinafter referred to as the EP side). The CPU and Orin1, Orin2, Orin3, and Orin4 form a master and four-slave computing architecture. Orin adopts a 2:2 fault tolerance method, that is, Orin1 and Orin3 are mutually redundant, and Orin2 and Orin4 are mutually redundant. When Orin1 computing equipment fails, Orin3 will take over in real time; when Orin2 computing equipment fails, Orin4 will take over in real time; thus Orin computing fault tolerance can be achieved. If there is a hardware failure in the CPU computing unit, the decision-making planning and control algorithms will not be able to run, the communication between Orin and the CPU will also be interrupted, and the system will experience serious errors. The MCU will monitor whether the CPU hardware fails in real time. Once the CPU fails, the MCU will Executing the control command, such as outputting the control command of pulling over to the chassis, can avoid the safety accident caused by the CPU failure, thereby realizing the fault tolerance of the CPU computing unit.

Communication layer fault-tolerant design, this system includes 2 sets of communication networks, CPU and Orin1, Orin2, Orin3, Orin4 are all connected to PCIESwitch, CPU and Orin1, Orin2, Orin3, Orin4 use PCIE3.0 communication, CPU and Orin1, Orin2, Orin3 and Orin4 are also connected to the Gigabit Ethernet switch at the same time. The communication between the CPU and Orin1, Orin2, Orin3, and Orin4 uses Gigabit Ethernet communication. When the PCIESwitch or PCIE3.0 communication link fails, it can be switched to Gigabit Ethernet Communication link, so as to realize the fault-tolerant communication between CPU and Orin. Since the PCIE3.0 communication bandwidth is higher than that of Gigabit Ethernet, Gigabit Ethernet is mainly used as a backup network. The Ethernet is interconnected, and the MCU is also connected to the Gigabit switch.

Data IO layer fault-tolerant design, the whole system is connected to a variety of sensing devices, among which three types of main sensing data are fault-tolerant, the first type is GMSL2 vehicle camera data, which is directly output to Orin, and does not need to be transferred through the CPU; the second type It is laser radar data, which can also be directly output to Orin; the third type is millimeter-wave radar data, which is read by MCU through the chassis CAN port, then sent to CPU, and then output to Orin through PCIE3.0, in order to ensure system fault tolerance Real-time, the same camera, lidar, millimeter-wave radar data need to be connected to two Orins synchronously to achieve fault tolerance. This system can access up to 16 channels of camera data. The first 8 channels of camera data are sent to Orin1 and Orin3 at the same time. Orin1 and Orin3 can realize real-time synchronization of camera data. When Orin1 fails, Orin3 can take over immediately; the last 8 Orin2 and Orin4 can be connected to Orin2 and Orin4 at the same time. Orin2 and Orin4 can realize real-time synchronization of camera data. When Orin2 fails, Orin4 can take over immediately; and so on, if the system is connected to 4 laser radars and 4 millimeter wave radars, then Orin1 and Orin3 read the data of the first 2 channels of lidar and the first 2 channels of millimeter-wave radar at the same time, and Orin2 and Orin4 simultaneously read the data of the last 2 channels of laser radar and the last 2 channels of millimeter-wave radar, so that the data input of the three main sensing devices can be To achieve synchronization, there is no need to wait for a certain Orin to fail before reading the sensing data, avoiding the delay overhead caused by switching data reading, thus ensuring the data IO fault tolerance of the system.

The fault-tolerant design of the algorithm layer builds a high-fault-tolerant vehicle-mounted software system and realizes fault-tolerance in the application algorithm. This system is transformed based on the autopilot framework Autoware.Universe to realize the fault-tolerant function of the algorithm.

Software and hardware collaborative fault-tolerant design, to achieve the overall fault-tolerant system, the design process is as follows:

The MCU monitors the hardware status of the five computing chips Orin1, Orin2, Orin3, Orin4 and CPU in real time;

If the 2:2 fault-tolerant mode is selected, and the MCU monitors that the CPU or two or more Orins are faulty, the MCU will execute the pull-over parking control command to the chassis; the MCU will execute the pull-over parking control command to the chassis, and the automatic driving level will be reduced from L4 to L1. Hardware level fault tolerance. When the CPU has no faults or the number of Orin faults is not more than 2, the highly fault-tolerant autonomous driving software system based on Autoware.Universe is deployed to run on the CPU+4*Orin platform, and realizes cooperative fault tolerance with the hardware system, and the software system realizes upper-level algorithm-level fault tolerance , the underlying hardware provides fault tolerance for computing, communication, and data IO, and the level of autonomous driving is not downgraded, and L4 is still achieved.

If the 3:1 fault-tolerant mode is selected, only Orin1 and Orin3 are mutually redundant, and important perception algorithms such as forward camera detection and main lidar detection are running. When the MCU monitors that the CPU or Orin1 and Orin3 fail at the same time, the MCU will perform a side stop. The control command is given to the chassis, and the automatic driving level is lowered from L4 to L1 to achieve hardware-level fault tolerance. When the CPU is fault-free and Orin1 and Orin3 are different, but one or both of Orin2 and Orin4 are faulty, the automatic driving level is changed from L4 Reduced to L2; when the CPU is fault-free and Orin1 and Orin3 are different, and Orin2 and Orin4 are also fault-free, the high-fault-tolerant autopilot software system based on Autoware.Universe is deployed to run on the CPU+4*Orin platform, and the software system realizes The upper-level algorithm level is fault-tolerant, and the underlying hardware provides fault-tolerance for computing, communication, and data IO. The level of autonomous driving is not downgraded, and L4 is still achieved;

The method embodiments provided in the embodiments of the present application may be executed in mobile terminals, computer terminals or similar computing devices. Taking running on a mobile terminal as an example, FIG. 4 is a block diagram of a hardware structure of a mobile terminal according to a driving control method according to an embodiment of the present application. As shown in FIG. 4, the mobile terminal may include one or more (only one is shown in FIG. 4) processors 402 (processors 402 may include but not limited to processing devices such as microprocessor MCU or programmable logic device FPGA) and a memory 404 for storing data, wherein the above-mentioned mobile terminal may also include a transmission device 406 and an input and output device 408 for communication functions. Those of ordinary skill in the art can understand that the structure shown in FIG. 4 is only for illustration, and it does not limit the structure of the above-mentioned mobile terminal. For example, the mobile terminal may also include more or fewer components than those shown in FIG. 4, or have a different configuration from that shown in FIG.

The memory 404 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the driving control method in the embodiment of the present application. The processor 402 executes various A functional application and data processing, that is, to realize the above-mentioned method. The memory 404 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 404 may further include a memory that is remotely located relative to the processor 402, and these remote memories may be connected to the mobile terminal through a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

Transmission device 406 is used to receive or transmit data via a network. The specific example of the above network may include a wireless network provided by the communication provider of the mobile terminal. In an example, the transmission device 406 includes a network interface controller (NIC for short), which can be connected to other network devices through a base station so as to communicate with the Internet. In an example, the transmission device 406 may be a radio frequency (Radio Frequency, RF for short) module, which is used to communicate with the Internet in a wireless manner.

In this embodiment, a driving control method is provided. FIG. 5 is a flow chart of the driving control method according to the embodiment of the present application. It is applied to the first processor in the driving control system. As shown in FIG. 5 , the flow Including the following steps:

Step S502, in the case that the driving device needs to be controlled according to the target driving control scene, determine a target algorithm corresponding to the target driving control scene, wherein the target algorithm is used to determine the driving environment of the driving device;

Step S504, selecting a master processor and a slave processor from multiple second processors deployed in the driving control system, wherein the first processor is respectively connected to each second processor, and the master processor is used to run the target algorithm , output the target calculation result, the target calculation result is used to characterize the driving environment of the driving equipment, and the slave processor is used to replace the main processor to run the target algorithm and output the target calculation result corresponding to the target algorithm when the main processor is in a fault state ;

Step S506, obtaining the target operation result.

Step S508, generating a driving control instruction to the driving equipment according to the target calculation result, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment.

Through the above steps, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each second processor, and when it is necessary to control the driving of the driving equipment according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

In the embodiment provided in step S502, the driving control scene and the target algorithm have a corresponding relationship, and different driving control scenes need to use different target algorithms. For example, when the driving control scene is to perform lane keeping control on the vehicle, then The required target algorithm can be an image recognition algorithm to identify the lane line of the lane. When the driving control scene assists the braking scene, the required target algorithm can be an image recognition algorithm and a perception algorithm, that is, the driving environment can be recognized through the image recognition algorithm In the object information, the relative positional relationship between the object and the ego vehicle is sensed through the perception algorithm, which is not limited in this scheme.

In the embodiment provided in step S504, the selection of the main processor and the slave processor from the second processor may be obtained by filtering the second processor according to the target computing power value required to run the target algorithm, for example, driving 4 second processors with the same computing power are deployed in the control system, and the computing power of each second processor is 225 TOPS. When the target computing power required by the target algorithm is lower than 550 TOPS (that is, 2 Orin (the The sum of the computing power of two processors), the entire autonomous driving algorithm stack can meet the computing power requirements by running on two Orins, and can form two Orin computing units among the four Orin computing units with the other two Orin computing units One-to-one computing redundancy, that is, two Orins are used as the main processor and the other two are used as slave processors. When any Orin computing device fails, its redundant Orin can take over in real time, so that Orin computing can be fault-tolerant.

Optionally, in this embodiment, when there are multiple target algorithms, selecting the master processor and the slave processor from the second processor may also be based on the target computing power value required to run the target algorithm and the target algorithm The corresponding algorithm importance is obtained by screening the second processor. For example, four second processors with the same computing power are deployed in the driving control system, and the computing power of each second processor is 225 TOPS. When the target algorithm The sum of the target computing power value is larger than 550TOPS (that is, the sum of the computing power of 2 Orins), and the entire algorithm stack needs to run on 3 Orins to meet the computing power requirements. The 4 Orins can be divided into There are 2 parts, of which 3 Orins are used as the main computing unit, and 1 Orin is used as the fault-tolerant computing unit. It forms a one-to-one redundancy with one Orin in the main computing unit, and mainly deals with perception algorithms with high importance of algorithms, such as for For the perception algorithm of the front camera, this fault-tolerant method only implements fault-tolerant processing for key algorithms. These key algorithms all run on one Orin. When the Orin computing device fails, its redundant Orin can take over in real time, so that Orin computing can be fault-tolerant. .

In the embodiment provided in step S506, when the target algorithm corresponding to the target calculation result runs on the master processor and the slave processor, at this time, the communication status of the target algorithm on the master processor and the slave processor can be passed Perform analysis to determine whether to obtain the first operation result output after the main processor runs the target algorithm as the target operation result, or obtain the second operation result output after the slave processor runs the target algorithm as the target operation result, for example, algorithm A and Algorithm B is the same target algorithm. Algorithm B runs on the main processor, and Algorithm A runs on the slave processor. You can judge the communication status of Algorithm B and Algorithm C respectively, such as judging the communication frequency status and communication status of Algorithm B. Time state, if it is normal, and algorithm B is a positioning, planning or sensor algorithm, it is necessary to further judge whether the data output by algorithm B is normal, if the communication status and data of algorithm B are normal, then select algorithm B run by the main processor The output result of algorithm A is used as the target operation result, otherwise, the output of algorithm A running on the processor is selected as the target operation result; if algorithm B is not a positioning, planning or sensor algorithm, there is no need to judge whether the data output by algorithm B is abnormal; if Algorithm B is a positioning, planning or sensor algorithm, then the operation to determine whether the output of algorithm B is abnormal is: call the positioning error monitoring algorithm, planning error monitoring algorithm or sensor error monitoring algorithm to perform positioning error, planning If the error is normal, it means that the data output by Algorithm B is normal, otherwise it is abnormal.

In the example provided in step S508, the driving control command may be determined from the corresponding calculation result and the driving control command to determine the driving control quality corresponding to the target calculation result, or it may also be through the corresponding driving control scene The decision algorithm is obtained by operating the target operation result, which is not limited in this solution.

As an optional embodiment, the main processor and the slave processor used to run the target algorithm are selected from multiple second processors, including:

Determine the target computing power value required to run the target algorithm;

The main processor and the slave processor are selected from the plurality of second processors according to the target computing power value.

Optionally, in this embodiment, the target computing power value is a computing power value of a processor used to run the target algorithm when the target algorithm is running.

Optionally, in this embodiment, the number of target algorithms may be one or more. When there is one target algorithm, the processor whose computing power value is greater than or equal to the target computing power value in the second processor may be used as The main processor, and the processor whose computing power value is greater than or equal to the target computing power value among the processors other than the main processor in the second processor is used as a slave processor; when the number of target algorithms is multiple, the One or more of the second processors whose total computing power value is greater than or equal to the sum of the target computing power values of multiple target algorithms is used as the main processor, and the other processors are used as slave processors; or when the target algorithm When the number is multiple, it is also possible to select a candidate algorithm whose algorithm importance is greater than the target importance from multiple target algorithms, and select two computing power values greater than or equal to the target of the latter candidate algorithm from multiple second processors. The processors of the computing power are used as the main processor and the slave processor respectively to implement the fault-tolerant operation of the candidate algorithm, and select the second processor from the processors other than the main processor or the slave processor for running Algorithms other than candidate algorithms are selected from multiple target algorithms. This solution does not limit the method of selecting the master processor and slave processor from multiple second processors according to the target computing power value.

As an optional embodiment, the main processor and the slave processor are selected from multiple second processors according to the target computing power value, including:

When the target computing power value is less than or equal to the preset computing power value, the target processor is selected from the multiple second processors as the main processor, wherein the target processor is the computing power of the multiple second processors For a processor with a value greater than or equal to the target computing power value, the preset computing power value is half of the sum of the computing power values of multiple second processors;

A processor other than the master processor among the plurality of second processors is determined as a slave processor.

Optionally, in this embodiment, the target processor may be a randomly selected processor among multiple second processors whose computing power value is greater than or equal to the target computing power value, or the target processor may also be For the multiple processors selected from the multiple second processors, the computing power sum of the multiple target processors is greater than or equal to the target computing power value.

As an optional embodiment, the main processor and the slave processor are selected from multiple second processors according to the target computing power value, including:

When the target computing power value is greater than the preset computing power value, determine the algorithm importance of each target algorithm among the multiple target algorithms, where the algorithm importance is used to indicate the impact of the target calculation result corresponding to the target algorithm on the driving control instruction impact on accuracy;

Determine candidate algorithms from multiple target algorithms according to the algorithmic importance of each target algorithm;

According to the computing power values and candidate algorithms of multiple second processors, the main processor and the slave processor are screened out from the second processors, wherein the computing power values of the main processor and the slave processor are both greater than or equal to the running candidate The computing power value required by the algorithm.

Optionally, in this embodiment, due to the different functions of different algorithms, the target results output by different algorithms have different influences on the accuracy of driving control instructions, and thus the importance of the algorithms is also different, for example, In an assisted driving control operation, an image recognition algorithm, an image definition processing algorithm, and a perception algorithm are required. Among these three algorithms, when the perception algorithm operates abnormally, the accuracy of the assisted driving control command output to the driving control system The influence of the image recognition algorithm and the image definition processing algorithm on the accuracy of the output assisted driving control instruction is lower than that of the perception algorithm on the accuracy of the assisted driving control instruction. Therefore, the perception Algorithms are more important than image recognition algorithms and image definition processing algorithms.

Optionally, the candidate algorithm may be one or more algorithms among multiple target algorithms whose algorithm importance is greater than or equal to the target importance, which is not limited in this solution.

Optionally, in this embodiment, the master processor and the slave processor are processors for running the candidate algorithm, because the influence degree of the candidate algorithm on the accuracy of the output driving control command is greater than other algorithms in the multiple target algorithms , so the candidate algorithm is redundantly set to avoid the abnormal output of the driving control command caused by the abnormality of the candidate algorithm.

As an optional embodiment, the candidate algorithm is determined from multiple target algorithms according to the algorithm importance of each target algorithm, including:

An algorithm whose algorithm importance is greater than the target importance is selected from multiple target algorithms as a candidate algorithm, wherein the target importance is determined according to the algorithm importance of the multiple target algorithms.

As an optional embodiment, according to the computing power values and candidate algorithms of multiple second processors, the main processor and the slave processor are selected from the second processors, including:

Select a processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm from a plurality of second processors as the main processor, and select a processor other than the main processor included in the plurality of second processors Filter out the processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm as the slave processor.

Optionally, in this embodiment, the computing power of the master processor and the slave processor may be the same or different, which is not limited in this solution.

As an optional embodiment, before obtaining the target operation result, the method further includes:

Add the first algorithm startup information and the second algorithm startup information to the startup file, wherein the first algorithm startup information is the information of the algorithm that needs to be run on the master processor, and the second algorithm startup information is the information that needs to be run on the slave processor Information about the running algorithm, the startup file is used to run the algorithm indicated by the algorithm startup information on the processor indicated by the algorithm startup information.

Optionally, in this embodiment, the startup file is used to start the corresponding algorithm according to the algorithm execution sequence indicated by the driving control scene, the startup file can be but not limited to a launch (startup) file, and the target algorithm is combined through the launch file , so as to form an algorithm pipeline (pipeline), so that the target algorithm runs on the corresponding second processor.

Optionally, in this embodiment, the startup information may be, but not limited to, identification information used to indicate the target algorithm, or may also be used to indicate the second processor used to run the target algorithm (the Only one target algorithm is stored), or it can also be the target algorithm to be run, which is not limited in this solution.

As an optional embodiment, before obtaining the target operation result, the method includes:

Determine the reference data subscription relationship between any two target algorithms, where the reference data subscription relationship is used to indicate that multiple target algorithms are running, and the target algorithm that runs later is used for the output of the previous target algorithm The result is data operation;

According to the reference data subscription relationship, modify the target data subscription relationship between the target algorithms running on the master processor and the slave processor respectively.

Optionally, in this embodiment, in a driving control scenario, in order to meet its functional requirements, multiple target algorithms with different functions need to be used, and the algorithm pipeline is formed by combining multiple target algorithms. In the algorithm running sequence specified by the algorithm pipeline, there may be a subscription relationship between the calculation results between multiple algorithms, that is, in the algorithm running sequence specified by the algorithm pipeline, there is a data subscription between the target algorithm that runs before and the target algorithm that runs after relationship, that is, the target calculation result output by the target algorithm that runs earlier needs to be passed to the target algorithm that runs later, and the target algorithm that runs later needs to perform calculations based on the results output by the target algorithm that runs earlier, so as to output the calculation results, such as , an image definition adjustment algorithm, an image recognition algorithm, and a perception algorithm are required in an assisted driving scene. These three algorithms form an algorithm pipeline. The image output by the image definition adjustment algorithm will be passed to the image recognition algorithm. Recognition is carried out to identify the objects in the image, and the perception algorithm operates on the objects in the image to obtain the positional relationship of the objects.

Optionally, in this embodiment, the master processor and the slave processor corresponding to different target algorithms need to modify the data subscription relationship between the algorithms they run, for example, target algorithm 1 and target algorithm 2 are of the same type target algorithm, target algorithm 1 runs on master processor A, target algorithm 2 runs on slave processor A, target algorithm 3 needs to subscribe to the calculation results of target algorithm 1 or target algorithm 2, and target algorithm 3 runs on the master processor On B, the subscription of the target algorithm 3 on the master processor B is modified to subscribe to both the target algorithm 1 on the master processor A and the target algorithm 2 on the slave processor A.

As an optional embodiment, after modifying the data subscription relationship between the master processor and the slave processor for running each target algorithm according to the data subscription relationship, the method further includes:

When the target data subscription relationship indicates that the first algorithm subscribes to the operation result of the second algorithm, and the second algorithm is configured to run on the master processor and the slave processor, call the communication monitoring algorithm to monitor the operation of the master processor The target communication state of the second algorithm, wherein the target algorithm includes the first algorithm and the second algorithm, and the target communication state includes the communication frequency state and the communication time state of the second algorithm;

According to the communication state of the target, the operation result for transmission to the first algorithm is selected from the first operation result and the second operation result, wherein the first operation result is the result output by the second algorithm run by the main processor, and the second operation The result is the output from the processor running the second algorithm.

Optionally, in this embodiment, when the target communication state indicates that the communication of the second algorithm running on the main processor is normal, the first calculation result is transmitted to the first algorithm, and when the target communication state indicates that the second algorithm running on the main processor is When the communication status of the running second algorithm is abnormal, the second calculation result is transmitted to the first algorithm.

As an optional embodiment, according to the target communication state, the calculation result for transmission to the first algorithm is selected from the first calculation result and the second calculation result, including:

In the case where the target communication state indicates that the second algorithm running on the main processor is in an abnormal communication state, determining that the second operation result is an operation result for transmission to the first algorithm;

When the target communication state indicates that the second algorithm running on the main processor is in a normal communication state, call the error monitoring algorithm to monitor the error value of the first operation result; when the error value of the first operation result is less than or equal to the target error value Under normal circumstances, it is determined that the first operation result is the operation result for transmission to the first algorithm; in the case where the error value of the first operation result is greater than the target error value, it is determined that the second operation result is the operation result for transmission to the first algorithm Operation result.

Optionally, in this embodiment, the error monitoring algorithm may include, but is not limited to, a positioning error monitoring algorithm, a planning error monitoring algorithm, a sensor error monitoring algorithm, etc., which are not limited in this solution.

Optionally, in this embodiment, whether it is necessary to call the error monitoring algorithm to monitor the calculation results can be determined according to the algorithm type of the second algorithm, for example, when the second algorithm is a positioning algorithm, a planning algorithm or a sensor algorithm , it is necessary to call the error monitoring algorithm to monitor the operation results.

The driving control method in the embodiment of the present application realizes fault tolerance in the application algorithm. This system is transformed based on the automatic driving framework Autoware.Universe to realize the driving control method with algorithmic fault tolerance function and ensure the reliability of the driving control method. Figure 6 is based on An optional driving control flow chart of the embodiment of the present application, as shown in Figure 6, at least includes the following steps:

Step S601, select the open source autopilot framework Autoware.Universe as the base system of the highly fault-tolerant in-vehicle software, its bottom layer is based on ROS2+DDS middleware, DDS is used as the communication middleware, ROS2 is used as the upper layer programming interface, and the upper layer has automatic driving perception based on ROS2 , Fusion, decision planning, control and other algorithms, according to the needs of your own application scenarios, you can select the corresponding algorithm components from the framework to build the application algorithm Pipline;

Step S602, according to the application scenario that needs to be built, based on the largest launch file logging_simulator.launch.xml in the Autoware.Universe framework as a template, the algorithm is cut or added to form a new launch file and the algorithm pipeline. The implementation process still uses the launch logging_simulator.launch.xml without any changes, so that it can run through the CPU first;

Step S603, divide the algorithm tasks, divide all the algorithms contained in the launch logging_simulator.launch.xml into three parts, and divide the logical processing algorithms such as state monitoring and control to run on the CPU; Orin1 processes the perception algorithm; Orin3 processes the rest For algorithms that require GPU and AI computing, such as fusion and positioning algorithms, modify the logging_simulator.launch.xml file so that the corresponding divided algorithms run on the CPU, Orin1 and Orin2 respectively;

Step S604, copy the algorithms of Orin1 and Orin2 to run on Orin3 and Orin4 respectively, add the copied algorithm nodes to the logging_simulator.launch.xml file, Orin1 and Orin3 will form an algorithm fault-tolerant, and both run the perception algorithm; Orin2 and Orin4 form an algorithm Fault tolerance, all run algorithms such as fusion and positioning;

Step S605, modify the input-output relationship of the algorithm, that is, modify the subscription-publishing relationship of the algorithm. The perception algorithm on Orin1 is changed from only subscribing to the output results of the positioning algorithm on Orin2 to subscribing to the output results of the positioning algorithms on Orin2 and Orin4; the perception algorithm on Orin1 The output of the algorithm is changed from the fusion algorithm released only to Orin2 to the fusion algorithm released to Orin2 and Orin4, and the input-output relationship of other algorithms is similarly modified;

Step S606, judge the results of subscribing to two same algorithms, and select one of them to execute. A two-level judging mechanism is adopted, as follows: Judging by the communication layer, judging whether the communication between the positioning algorithms of Orin2 and Orin4 is normal, the sensing algorithm on Orin1 will call the status monitoring algorithm topic_static_monitor running on the CPU Package), which judges the communication frequency and time status of Orin2's positioning algorithm. If it is normal, and because it is a positioning algorithm, it is necessary to further judge whether the data output by Orin2's positioning algorithm is normal. If the Orin2 positioning algorithm, its communication and data are both If it is normal, select the output of the positioning algorithm running on Orin2; otherwise, if the positioning algorithm running on Orin2 is abnormal, then select the output of the positioning algorithm running on the fault-tolerant node Orin4; for data layer judgment, the perception algorithm of Orin1 calls the positioning error monitoring algorithm localization_error_monitor (Algorithm package that comes with the Autoware.Universe framework) Judging the positioning error of the data output by Orin2's positioning algorithm, if the error is normal, it means that the data output by Orin2's positioning algorithm is normal, otherwise it is abnormal;

Step S607, complete the above operations, the overall algorithm pipeline will start based on a launch file, and complete the operation on CPU+4*Orin, run the perception algorithm on Orin1 and Orin3 to achieve fault tolerance, and run fusion and positioning algorithms on Orin2 and Orin4 to achieve fault tolerance , the overall Autoware.Universe framework realizes the algorithm fault-tolerant function, so as to obtain the driving control instructions of the high-speed driving equipment.

In this embodiment, a driving control method is provided. FIG. 7 is the second flow chart of the driving control method according to the embodiment of the present application, which is applied to the third processor in the driving control system. As shown in FIG. 7 , the flow Including the following steps:

Step S702, monitoring the first running state of the first processor deployed in the driving control system and the second running state of multiple second processors, wherein the first processor is respectively connected to each second processor, and the third The processor is connected to the first processor and each of the second processors, and the first processor is used to determine the target algorithm corresponding to the target driving control scene when it is necessary to control the driving of the driving device according to the target driving control scene, and learn from multiple The main processor and the slave processor used to run the target algorithm are selected from the second processor, and the target algorithm is used to determine the driving environment of the driving equipment; the main processor is used to run the target algorithm and output the target calculation result, and the target calculation result is used for It is used to characterize the driving environment of the driving equipment; the slave processor is used to replace the main processor to run the target algorithm when the main processor is in a fault state, and output the target calculation result corresponding to the target algorithm; the first processor is also used to obtain the target Calculate the result of the calculation, and generate a driving control command for the driving device according to the target calculation result, and the driving control command is used to indicate the driving state of the driving device in the driving environment;

Step S704, according to the first running state and the second running state, determine the driving automation level of the driving equipment.

Through the above steps, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each second processor, and when it is necessary to control the driving of the driving equipment according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

Optionally, in this embodiment, the driving automation levels include: L0 level (pure manual driving), L1 level (assisted driving), L2 level (partially automated driving), L3 level (conditional automated driving), L4 level (highly autonomous driving), L5 level (fully autonomous driving).

As an optional embodiment, determining the driving automation level of the driving device according to the first operating state and the second operating state includes:

When the master processor and the slave processor are determined by the first method, and if the first preset condition is satisfied, the driving automation level of the driving equipment is determined as the assisted driving level, wherein the first preset condition includes the following One: the first operating state indicates that the first processor is in a fault state, and the second operating state indicates that the number of processors in a fault state among the plurality of second processors is greater than or equal to half of the total number of the plurality of second processors, wherein the first One method includes: when the target computing power value required to run the target algorithm is less than or equal to the preset computing power value, select the target processor from the multiple second processors as the main processor, and use the multiple second processors Among the processors, the processors other than the main processor are determined as slave processors, the target processor is a processor whose computing power value is greater than or equal to the target computing power value among multiple second processors, and the preset computing power value is multiple Half of the sum of the computing power of the second processor;

When the master processor and the slave processor are determined by the first method and the second preset condition is met, determine that the driving automation level of the driving equipment is a high automation level, wherein the second preset condition includes the following One: the first operating state indicates that the first processor is in a normal state; the second operating state indicates that the number of processors in a fault state among the plurality of second processors is less than half of the total number of the plurality of second processors.

As an optional embodiment, after determining the driving automation level of the traveling device according to the first operating state and the second operating state, the method further includes:

When the master processor and the slave processor are determined through the second method and the third preset condition is met, determine that the driving automation level of the driving equipment is the assisted driving level, wherein the third preset condition includes the following One: The first operating state is used to indicate that the first processor is in an abnormal state, and the second operating state is used to indicate that the master processor and the slave processor fail at the same time, wherein the second mode is to run the target algorithm required by the target algorithm When the power value is greater than the preset computing power value, determine the algorithm importance of each target algorithm among the multiple target algorithms, where the algorithm importance is used to indicate the accuracy of the target calculation result corresponding to the target algorithm to the driving control instruction. According to the algorithm importance of each target algorithm, candidate algorithms are determined from multiple target algorithms, and based on the computing power values and candidate algorithms of multiple second processors, the main processor and The computing power of the slave processor, master processor and slave processor is greater than or equal to the computing power value of the candidate algorithm;

When the master processor and the slave processor are determined by the second method, and in the case that the fourth preset condition is satisfied, the driving automation level of the driving equipment is determined to be a partially automated driving level, wherein the fourth preset condition includes: The first operating state is used to indicate that the first processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processor are not in a fault state at the same time, the third processor is in a fault state, and the third processor is a plurality of Processors other than the master processor and the slave processor in the second processor;

When the master processor and the slave processor are determined by the second method and the fifth preset condition is satisfied, the driving automation level of the driving equipment is determined to be a highly automated driving level, wherein the fifth preset condition includes: The first operating state is used to indicate that the first processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processors are not in a failure state at the same time, and the third processor is in a normal state.

As an optional embodiment, after determining the driving automation level of the traveling device according to the first operating state and the second operating state, the method further includes:

When the driving automation level is lower than the target level, a control instruction is sent to the driving equipment, and the control instruction is used to instruct the driving equipment to pull over and stop.

Optionally, in this embodiment, after the driving automation level is determined, the control instruction corresponding to the current driving automation level can also be determined from the corresponding driving automation level and control instructions, and sent to the chassis control The system instructs the chassis control system to perform safety control operations on the driving state of the vehicle (including but not limited to deceleration, parking, alarm, etc.).

Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is Better implementation. Based on this understanding, the essence of the technical solution of this application or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products are stored in a storage medium (such as ROM/RAM, disk, CD-ROM), including several instructions to enable a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of various embodiments of the present application.

In this embodiment, a travel control device is also provided, which is used to implement the above embodiments and preferred implementation modes, and what has already been described will not be repeated. As used below, the term "module" may be a combination of software and/or hardware that realizes a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementations in hardware, or a combination of software and hardware are also possible and contemplated.

Fig. 8 is a structural block diagram 1 of a driving control device according to an embodiment of the present application. As shown in Fig. 8 , the device includes: a first determining module 82, which is used to control the driving of the driving equipment according to the target driving control scene, Determine the target algorithm corresponding to the target driving control scenario, wherein the target algorithm is used to determine the driving environment of the driving device; the first screening module 84 is used to filter out the main processing from multiple second processors deployed in the driving control system The processor and the slave processor, wherein, the first processor is respectively connected with each second processor, the main processor is used to run the target algorithm, and output the target calculation result, the target calculation result is used to characterize the driving environment of the driving device, and the slave processor uses When the main processor is in a fault state, take over the main processor to run the target algorithm and output the target calculation result corresponding to the target algorithm; the acquisition module 86 is used to obtain the target calculation result; As a result, a driving control instruction for the driving equipment is generated, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment.

Based on the above content, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each second processor, and when it is necessary to control the driving of the driving device according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

Optionally, the first screening module includes: a determination unit, configured to determine a target computing power value required to run the target algorithm; a screening unit, configured to filter out the main processor from multiple second processors according to the target computing power value processor and slave processor.

Optionally, the screening unit is configured to: select the target processor from multiple second processors as the main processor when the target computing power value is less than or equal to the preset computing power value, wherein the target processor It is a processor whose computing power value is greater than or equal to the target computing power value among the multiple second processors, and the preset computing power value is half of the sum of the computing power values of multiple second processors; multiple second processors Processors other than the master processor are identified as slave processors.

Optionally, the screening unit is configured to: determine the algorithm importance of each target algorithm in multiple target algorithms when the target computing power value is greater than the preset computing power value, wherein the algorithm importance is used to indicate the target algorithm The impact of the corresponding target calculation results on the accuracy of the driving control command; determine the candidate algorithm from multiple target algorithms according to the algorithm importance of each target algorithm; , and select the master processor and the slave processor from the second processor, wherein the computing power values of the master processor and the slave processor are both greater than or equal to the computing power value required to run the candidate algorithm.

Optionally, the screening unit is configured to: select an algorithm whose algorithm importance is greater than the target importance from multiple target algorithms as a candidate algorithm, wherein the target importance is determined according to the algorithm importance of the multiple target algorithms.

Optionally, the screening unit is configured to: select a processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm from multiple second processors as the main processor, and select from multiple second processors including Among the processors other than the main processor, the processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm is selected as the slave processor.

Optionally, the device further includes: an adding module, configured to add the first algorithm startup information and the second algorithm startup information to the startup file before obtaining the target operation result, wherein the first algorithm startup information is required in the main processing The second algorithm startup information is information about the algorithm that needs to be run on the slave processor, and the startup file is used to run the algorithm indicated by the algorithm startup information on the processor indicated by the algorithm startup information.

Optionally, the device further includes: a second determining module, configured to determine a reference data subscription relationship between any two target algorithms before obtaining the target operation result, wherein the reference data subscription relationship is used to indicate that multiple target algorithms are in During the running process, the target algorithm running later is used to perform data calculations based on the calculation results output by the target algorithm running earlier; the modification module is used to modify the data running on the master processor and the slave processor according to the reference data subscription relationship. The target data subscription relationship between the target algorithms.

Optionally, the device further includes: a calling module, configured to indicate the first algorithm in the target data subscription relationship after modifying the data subscription relationship between the master processor and the slave processor for running each target algorithm according to the data subscription relationship. The operation result of the second algorithm is subscribed, and the second algorithm is configured to call the communication monitoring algorithm to monitor the target communication state of the second algorithm running on the main processor when the second algorithm is configured to run on the main processor and the slave processor, wherein , the target algorithm includes the first algorithm and the second algorithm, and the target communication state includes the communication frequency state and the communication time state of the second algorithm; the second screening module is used to select from the first operation result and the second operation result according to the target communication state The operation result for transmission to the first algorithm is filtered out, wherein the first operation result is the result output by the master processor running the second algorithm, and the second operation result is the result output by the slave processor running the second algorithm.

Optionally, the second screening module includes: a third determination unit, configured to determine that the second operation result is used for transmission to the The operation result of the first algorithm; the first processing unit is used to call the error monitoring algorithm to monitor the error value of the first operation result when the target communication state indicates that the second algorithm running on the main processor is in a normal communication state; When the error value of the first operation result is less than or equal to the target error value, determine the first operation result as the operation result for transmission to the first algorithm; when the error value of the first operation result is greater than the target error value, The second operation result is determined as the operation result for transmission to the first algorithm.

Fig. 9 is a structural block diagram II of a driving control device according to an embodiment of the present application. As shown in Fig. 9 , the device includes: a monitoring module 92 for monitoring the first operating state and the first processor deployed in the driving control system The second operating state of a plurality of second processors, wherein the first processor is respectively connected to each second processor, the third processor is connected to the first processor and each second processor, and the first processor uses To determine the target algorithm corresponding to the target driving control scenario when the driving equipment needs to be controlled according to the target driving control scenario, and select the main processor and the slave processing for running the target algorithm from a plurality of second processors The target algorithm is used to determine the driving environment of the driving equipment; the main processor is used to run the target algorithm and output the target calculation results, and the target calculation results are used to characterize the driving environment of the driving equipment; Under the circumstances, take over the main processor to run the target algorithm, and output the target calculation result corresponding to the target algorithm; the first processor is also used to obtain the target calculation result, and generate a driving control command for the driving equipment according to the target calculation result, and the driving control command It is used to indicate the driving state of the driving device in the driving environment; the third determination module 94 is used to determine the driving automation level of the driving device according to the first running state and the second running state.

Based on the above content, the driving control system includes a plurality of first processors and a plurality of second processors, the first processors are respectively connected to each second processor, and when it is necessary to control the driving of the driving device according to the target driving control scene , the first processor selects the main processor and the slave processor for running the target algorithm from a plurality of second processors, runs the target algorithm through the main processor and outputs the target operation result, and realizes the operation in the main processor through the slave processor. When the processor fails, it replaces the main processor to run the target algorithm, which realizes the self-redundancy of the computing power of the target algorithm in the driving control system. In the driving control system, the target algorithm is run by the second processor, and the first processor is used for The operation results of the target algorithm output by the second processor generate corresponding control instructions, so as to realize the self-redundancy of the processor used to run the target algorithm in the driving control system to ensure that the target algorithm can be run accurately and ensure driving control The driving control function of the system on the driving equipment, therefore, can solve the problem of low control efficiency of the driving control system on the driving equipment in the related art, and achieve the effect of improving the control efficiency of the driving control system on the driving equipment.

Optionally, the third determining module includes: a second processing unit, configured to determine the driving device’s driving device when the master processor and the slave processor are determined through the first method and meet the first preset condition. The driving automation level is an assisted driving level, wherein the first preset condition includes one of the following: the first operating state indicates that the first processor is in a fault state, and the second operating state indicates that processing among the plurality of second processors is in a fault state The number of processors is greater than or equal to half of the total number of multiple second processors, wherein the first method includes: when the target computing power value required to run the target algorithm is less than or equal to the preset computing power value, from multiple second processors The target processor is selected from the processors as the main processor, and the processors other than the main processor among the multiple second processors are determined as the slave processors, and the target processor is that the computing power value of the multiple second processors is greater than Or a processor equal to the target computing power value, the preset computing power value is half of the sum of the computing power of multiple second processors; Determined, and when the second preset condition is met, it is determined that the driving automation level of the driving device is a highly automated driving level, wherein the second preset condition includes one of the following: the first operating state indicates that the first processor is in the Normal state; the second operating state indicates that the number of processors in the fault state among the plurality of second processors is less than half of the total number of the plurality of second processors.

Optionally, the device further includes: a fourth determining module, configured to determine the driving automation level of the driving device according to the first operating state and the second operating state, after the master processor and the slave processor determine the driving automation level in the second manner. , and if the third preset condition is met, it is determined that the driving automation level of the driving device is the assisted driving level, wherein the third preset condition includes one of the following: the first operating state is used to indicate that the first processor is abnormal state, the second operating state is used to indicate that the master processor and the slave processor have failures at the same time, wherein the second way is to determine multiple Algorithm importance of each target algorithm in the target algorithm, where the algorithm importance is used to indicate the impact of the target calculation result corresponding to the target algorithm on the accuracy of the driving control command, according to the algorithm importance of each target algorithm from multiple The candidate algorithm is determined from the target algorithm, and the master processor and the slave processor are selected from the second processors according to the computing power values of multiple second processors and the candidate algorithms. The computing power of the master processor and the slave processor are both Greater than or equal to the computing power value of the candidate algorithm; the second determination module is used to determine the driving speed of the driving device when the main processor and the slave processor are determined by the second method and the fourth preset condition is met. The automation level is a partially automated driving level, wherein the fourth preset condition includes: the first operating state is used to indicate that the first processor is in a normal state, and the second operating state is used to indicate that the master processor and the slave processor are not in failure at the same time state, the third processor is in a fault state, and the third processor is a processor other than the main processor and the slave processor in a plurality of second processors; the sixth determination module is used for the master processor and the slave processor It is determined through the second method, and if the fifth preset condition is met, it is determined that the driving automation level of the driving equipment is a highly automated driving level, wherein the fifth preset condition includes: the first operating state is used to indicate the first One processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processor are not in a failure state at the same time, and the third processor is in a normal state.

Optionally, the device further includes: a sending module, configured to, after determining the driving automation level of the driving equipment according to the first operating state and the second operating state, send a control to the driving equipment when the driving automation level is lower than the target level Command, the control command is used to instruct the driving equipment to pull over and stop.

It should be noted that the above-mentioned modules can be realized by software or hardware. For the latter, it can be realized by the following methods, but not limited to this: the above-mentioned modules are all located in the same processor; or, the above-mentioned modules can be combined in any combination The forms of are located in different processors.

Embodiments of the present application also provide a computer-readable storage medium, in which a computer program is stored, wherein the computer program is set to execute the steps in any one of the above method embodiments when running.

In an exemplary embodiment, the above-mentioned computer-readable storage medium may include but not limited to: U disk, read-only memory (Read-Only Memory, ROM for short), random access memory (Random Access Memory, RAM for short) , mobile hard disk, magnetic disk or optical disk and other media that can store computer programs.

An embodiment of the present application also provides an electronic device, including a memory and a processor, where a computer program is stored in the memory, and the processor is configured to run the computer program to perform the steps in any one of the above method embodiments.

In an exemplary embodiment, the electronic device may further include a transmission device and an input and output device, wherein the transmission device is connected to the processor, and the input and output device is connected to the processor.

For specific examples in this embodiment, reference may be made to the examples described in the foregoing embodiments and exemplary implementation manners, and details will not be repeated here in this embodiment.

Obviously, those skilled in the art should understand that each module or each step of the above-mentioned application can be realized by a general-purpose computing device, and they can be concentrated on a single computing device, or distributed in a network composed of multiple computing devices In fact, they can be implemented in program code executable by a computing device, and thus, they can be stored in a storage device to be executed by a computing device, and in some cases, can be executed in an order different from that shown here. Or described steps, or they are fabricated into individual integrated circuit modules, or multiple modules or steps among them are fabricated into a single integrated circuit module for implementation. As such, the present application is not limited to any specific combination of hardware and software.

The above are only preferred embodiments of the present application, and are not intended to limit the present application. For those skilled in the art, there may be various modifications and changes in the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of this application shall be included within the scope of protection of this application.

Claims (20)

1. A driving control system, characterized in that it comprises: a first processor and a plurality of second processors, wherein the first processor is respectively connected to each second processor, wherein: The first processor is used to determine the target algorithm corresponding to the target driving control scene when the driving device needs to be controlled according to the target driving control scene, and select the target algorithm for running the target algorithm from a plurality of second processors. a master processor and a slave processor, wherein the target algorithm is used to determine the driving environment of the driving device; The main processor is used to run the target algorithm and output the target calculation result, wherein the target calculation result is used to represent the driving environment of the driving device; The slave processor is used to replace the main processor to run the target algorithm when the main processor is in a fault state, and output the target operation result corresponding to the target algorithm; The first processor is further configured to obtain a target calculation result, and generate a driving control instruction for the driving equipment according to the target calculation result, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment. 2. The driving control system according to claim 1, wherein the driving control system further comprises a third processor, wherein the third processor is respectively connected to the first processor and each second processor, The third processor is configured to monitor the first operating state of the first processor and the second operating state of the second processor, and determine the driving automation level of the driving device according to the first operating state and the second operating state. 3. The driving control system according to claim 1, wherein the first processor and each second processor are connected by multiple data transmission links, The first processor is configured to monitor the link state of the first data transmission link among the multiple data transmission links, and when the first data transmission link is in a failure state, select the data transmission link from the multiple data transmission links determining a second data transmission link other than the first data transmission link, and transmitting data between the second data transmission link and the second processor, wherein the first data transmission link is the first processor's current A link to transfer data to and from the second processor. 4. The driving control system according to claim 1, characterized in that the driving control system further comprises an environment sensing device, wherein the sensing device is respectively connected to each second processor, The perception device is used to collect the environmental information of the driving environment, and transmit the environmental information to the main processor and the slave processor synchronously. 5. A driving control method, characterized in that it is applied to the first processor in the driving control system, comprising: When it is necessary to control the driving of the driving device according to the target driving control scene, determine a target algorithm corresponding to the target driving control scene, wherein the target algorithm is used to determine the driving environment of the driving device; The main processor and the slave processor are selected from multiple second processors deployed in the driving control system, wherein the first processor is connected to each second processor respectively, and the main processor is used to run the target algorithm and output the target The calculation result, the target calculation result is used to represent the driving environment of the driving equipment, and the slave processor is used to replace the main processor to run the target algorithm and output the target calculation result corresponding to the target algorithm when the main processor is in a fault state; Obtain the target operation result; A driving control instruction to the driving equipment is generated according to the target calculation result, wherein the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment. 6. The driving control method according to claim 5, wherein the main processor and the slave processor for running the target algorithm are selected from a plurality of second processors, including: Determine the target computing power value required to run the target algorithm; The main processor and the slave processor are selected from the plurality of second processors according to the target computing power value. 7. The driving control method according to claim 6, wherein the main processor and the slave processor are screened out from a plurality of second processors according to the target computing power value, comprising: When the target computing power value is less than or equal to the preset computing power value, the target processor is selected from the multiple second processors as the main processor, wherein the target processor is the computing power of the multiple second processors For a processor with a value greater than or equal to the target computing power value, the preset computing power value is half of the sum of the computing power values of multiple second processors; A processor other than the master processor among the plurality of second processors is determined as a slave processor. 8. The driving control method according to claim 6, wherein the main processor and the slave processor are screened out from a plurality of second processors according to the target computing power value, comprising: When the target computing power value is greater than the preset computing power value, determine the algorithm importance of each target algorithm among the multiple target algorithms, where the algorithm importance is used to indicate the impact of the target calculation result corresponding to the target algorithm on the driving control instruction impact on accuracy; Determine candidate algorithms from multiple target algorithms according to the algorithmic importance of each target algorithm; According to the computing power values and candidate algorithms of multiple second processors, the main processor and the slave processor are screened out from the second processors, wherein the computing power values of the main processor and the slave processor are both greater than or equal to the running candidate The computing power value required by the algorithm. 9. The driving control method according to claim 8, wherein the candidate algorithm is determined from a plurality of target algorithms according to the algorithm importance of each target algorithm, including: An algorithm whose algorithm importance is greater than the target importance is selected from multiple target algorithms as a candidate algorithm, wherein the target importance is determined according to the algorithm importance of the multiple target algorithms. 10. The driving control method according to claim 8, characterized in that, according to the computing power values and candidate algorithms of a plurality of second processors, the main processor and the slave processor are selected from the second processors, including: Select a processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm from a plurality of second processors as the main processor, and select a processor other than the main processor included in the plurality of second processors Filter out the processor whose computing power value is greater than or equal to the computing power value of the candidate algorithm as the slave processor. 11. The driving control method according to claim 5, characterized in that, before obtaining the target calculation result, the method further comprises: Add the first algorithm startup information and the second algorithm startup information to the startup file, wherein the first algorithm startup information is the information of the algorithm that needs to be run on the master processor, and the second algorithm startup information is the information that needs to be run on the slave processor Information about the running algorithm, the startup file is used to run the algorithm indicated by the algorithm startup information on the processor indicated by the algorithm startup information. 12. The driving control method according to claim 5, characterized in that, before obtaining the target calculation result, the method further comprises: Determine the reference data subscription relationship between any two target algorithms, where the reference data subscription relationship is used to indicate that multiple target algorithms are running, and the target algorithm that runs later is used for the output of the previous target algorithm The result is data operation; According to the reference data subscription relationship, modify the target data subscription relationship between the target algorithms running on the master processor and the slave processor respectively. 13. The driving control method according to claim 12, characterized in that, after modifying the target data subscription relationship between the target algorithms running on the master processor and the slave processor according to the reference data subscription relationship, the method further include: When the target data subscription relationship indicates that the first algorithm subscribes to the operation result of the second algorithm, and the second algorithm is configured to run on the master processor and the slave processor, call the communication monitoring algorithm to monitor the operation of the master processor The target communication state of the second algorithm, wherein the target algorithm includes the first algorithm and the second algorithm, and the target communication state includes the communication frequency state and the communication time state of the second algorithm; According to the communication state of the target, the operation result for transmission to the first algorithm is selected from the first operation result and the second operation result, wherein the first operation result is the result output by the second algorithm run by the main processor, and the second operation The result is the output from the processor running the second algorithm. 14. The driving control method according to claim 13, characterized in that, according to the target communication state, the calculation result for transmission to the first algorithm is selected from the first calculation result and the second calculation result, including: In the case where the target communication state indicates that the second algorithm running on the main processor is in an abnormal communication state, determining that the second operation result is an operation result for transmission to the first algorithm; When the target communication state indicates that the second algorithm running on the main processor is in a normal communication state, call the error monitoring algorithm to monitor the error value of the first operation result; when the error value of the first operation result is less than or equal to the target error value Under normal circumstances, it is determined that the first operation result is the operation result for transmission to the first algorithm; in the case where the error value of the first operation result is greater than the target error value, it is determined that the second operation result is the operation result for transmission to the first algorithm Operation result. 15. A driving control method, characterized in that it is applied to a third processor in a driving control system, comprising: monitoring the first operating state of the first processor deployed in the driving control system and the second operating state of multiple second processors, wherein the first processor is respectively connected to each second processor, and the third processor is connected to the The first processor and each second processor are connected, and the first processor is used to determine the target algorithm corresponding to the target driving control scene when the driving equipment needs to be controlled according to the target driving control scene, and obtain the information from multiple second processors. The main processor and slave processor used to run the target algorithm are selected from the processors. The target algorithm is used to determine the driving environment of the driving equipment; the main processor is used to run the target algorithm and output the target calculation result, which is used to represent the driving environment. The driving environment of the device; the slave processor is used to replace the main processor to run the target algorithm when the main processor is in a fault state, and output the target calculation result corresponding to the target algorithm; the first processor is also used to obtain the target calculation result, And according to the target calculation result, a driving control instruction for the driving equipment is generated, and the driving control instruction is used to indicate the driving state of the driving equipment in the driving environment; Depending on the first operating state and the second operating state, a driving automation level of the vehicle is determined. 16. The driving control method according to claim 15, wherein determining the driving automation level of the driving device according to the first operating state and the second operating state comprises: When the master processor and the slave processor are determined by the first method, and if the first preset condition is satisfied, the driving automation level of the driving equipment is determined as the assisted driving level, wherein the first preset condition includes the following One: the first operating state indicates that the first processor is in a fault state, and the second operating state indicates that the number of processors in a fault state among the plurality of second processors is greater than or equal to half of the total number of the plurality of second processors, wherein the first One method includes: when the target computing power value required to run the target algorithm is less than or equal to the preset computing power value, select the target processor from the multiple second processors as the main processor, and use the multiple second processors Among the processors, the processors other than the main processor are determined as slave processors, the target processor is a processor whose computing power value is greater than or equal to the target computing power value among multiple second processors, and the preset computing power value is multiple Half of the sum of the computing power of the second processor; When the master processor and the slave processor are determined by the first method, and in the case of satisfying the second preset condition, determine that the driving automation level of the driving equipment is a highly automated driving level, wherein the second preset condition includes the following One: the first operating state indicates that the first processor is in a normal state; the second operating state indicates that the number of processors in a fault state among the plurality of second processors is less than half of the total number of the plurality of second processors. 17. The driving control method according to claim 15, wherein determining the driving automation level of the driving equipment according to the first operating state and the second operating state comprises: When the master processor and the slave processor are determined through the second method and the third preset condition is met, determine that the driving automation level of the driving equipment is the assisted driving level, wherein the third preset condition includes the following One: The first operating state is used to indicate that the first processor is in an abnormal state, and the second operating state is used to indicate that the master processor and the slave processor fail at the same time, wherein the second mode is to run the target algorithm required by the target algorithm When the power value is greater than the preset computing power value, determine the algorithm importance of each target algorithm among the multiple target algorithms, where the algorithm importance is used to indicate the accuracy of the target calculation result corresponding to the target algorithm to the driving control instruction. According to the algorithm importance of each target algorithm, candidate algorithms are determined from multiple target algorithms, and based on the computing power values and candidate algorithms of multiple second processors, the main processor and The computing power of the slave processor, master processor and slave processor is greater than or equal to the computing power value of the candidate algorithm; When the master processor and the slave processor are determined by the second method, and in the case that the fourth preset condition is satisfied, the driving automation level of the driving equipment is determined to be a partially automated driving level, wherein the fourth preset condition includes: The first operating state is used to indicate that the first processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processor are not in a fault state at the same time, the third processor is in a fault state, and the third processor is a plurality of Processors other than the master processor and the slave processor in the second processor; When the master processor and the slave processor are determined by the second method and the fifth preset condition is satisfied, the driving automation level of the driving equipment is determined to be a highly automated driving level, wherein the fifth preset condition includes: The first operating state is used to indicate that the first processor is in a normal state, the second operating state is used to indicate that the master processor and the slave processors are not in a failure state at the same time, and the third processor is in a normal state. 18. The driving control method according to claim 15, characterized in that, after determining the driving automation level of the driving device according to the first operating state and the second operating state, the method further comprises: When the driving automation level is lower than the target level, a control instruction is sent to the driving equipment, and the control instruction is used to instruct the driving equipment to pull over and stop. 19. A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed by a processor, the steps of the method in any one of claims 5 to 18 are implemented. 20. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and operable on the processor, characterized in that, when the processor executes the computer program, the method in any one of claims 5 to 18 is implemented A step of.

Images