CN103335853B - A kind of automatic driving vehicle Cognitive Aptitude Test system and method - Google Patents

Abstract

The invention discloses a system for testing the cognitive ability of an unmanned vehicle. The system includes an unmanned vehicle and a vehicle-mounted recording device, both of which are wired through a communication interface, wherein the vehicle-mounted recording device is used to input Cognitive ability test task, and receive and store the output data of the unmanned vehicle cognitive ability test; after the unmanned vehicle receives the test task through the test task input program, re-establish the serial port connection, open the corresponding program to output the test data to the vehicle In the recording device, the test task input and test data output of the unmanned vehicle cognitive ability test adopt a unified interface, parameters and data transmission format, but the specific format of the data content is different; the communication interface adopts the RS-232 interface.

Description

A system and method for testing the cognitive ability of an unmanned vehicle

technical field

The invention relates to the field of vehicle intelligent systems, in particular to a system and method for testing cognitive ability of unmanned vehicles.

Background technique

Unmanned vehicles integrate the functions of environment perception, cognition, decision planning and driving control, and can drive autonomously, safely and reliably in specific environments. With the development of unmanned vehicle technology in recent years, both academia and industry need to design a method for testing and evaluating the intelligence of unmanned vehicles.

On the basis of independent research on unmanned vehicles conducted by various research institutions and enterprises, the United States, Europe, and China respectively began to conduct organized testing under a unified test environment, test content, or evaluation criteria in 2004, 2006, and 2009. Hold driverless vehicle races. The main events include: DARPA in the United States held two Grand Challenges and one Urban Challenge in 2004, 2005 and 2007 respectively; Europe has organized ELROB competitions for six consecutive times since 2006; NSFC has held four consecutive China Challenges since 2009. "Smart Car Future Challenge" contest.

However, no matter whether it is an organized unmanned vehicle competition at home or abroad, or Google's self-driving car self-test, although these events and tests stipulate some corresponding test rules, these rules are still not comprehensive, systematic and objective. reflect the level of cognitive ability of unmanned vehicles; although various research institutes also have relevant environments and methods designed to test their own unmanned vehicles, but these designs are based on their own unmanned vehicles. These tests are often non-standard and unsystematic.

Contents of the invention

In order to comprehensively, systematically and objectively test and evaluate the cognitive ability of unmanned vehicles, so that the tested unmanned vehicles have basic autonomous driving ability on urban roads, intercity roads and rural roads, the present invention proposes an unmanned A system and method for testing the cognitive ability of a human-driven vehicle.

According to one aspect of the present invention, an unmanned vehicle cognitive ability test system is proposed, the system includes an unmanned vehicle and a vehicle-mounted recording device, the two are wired or wirelessly connected through a communication interface, wherein the vehicle-mounted recording device uses It is used to input cognitive ability test tasks to unmanned vehicles, and receive and store the output data of unmanned vehicle cognitive ability tests; after unmanned vehicles receive test tasks through the test task input program, re-establish the serial port connection and start The corresponding program outputs test data to the on-board recording device. The test task input and test data output of the driverless vehicle cognitive ability test adopt a unified interface, parameters and data transmission format, but the specific format of the data content is different. Among them, the test tasks include Basic Cognitive Ability Test Items, Advanced Cognitive Ability Test Items and Comprehensive Cognitive Ability Test Items. .

According to another aspect of the present invention, a method for testing the cognitive ability of an unmanned vehicle is also proposed. The method includes the steps: the on-board recording device is used to input cognitive ability test tasks to the unmanned vehicle, and receives and stores the unmanned vehicle. The output data of the human-driven vehicle cognitive ability test; after the unmanned vehicle receives the test task through the test task input program, it re-establishes the serial port connection with the on-board recording device, and starts the corresponding program to output the test data to the on-board recording device. The test task input and test data output of the driving vehicle cognitive ability test adopt a unified interface, parameters and data transmission format, but the specific format of the data content is different. Among them, the test task includes basic cognitive ability test items, advanced cognitive ability test items and comprehensive cognitive ability test items. .

The invention can verify the safety, reliability, stability and intelligence of the operation of the unmanned vehicle by testing and evaluating the unmanned vehicle in a standardized test environment.

Description of drawings

Fig. 1 is an overall block diagram of the driverless vehicle cognitive ability testing system and method of the present invention;

Fig. 2 is a schematic diagram of the I-test scheme;

Fig. 3 is the scheme of M-test;

Figure 4 is a schematic diagram of an example of an M-test scenario;

Fig. 5 is a schematic diagram of lane keeping driving trajectory;

Figure 6 is a schematic diagram of parking at the intersection parking line;

Figure 7 is a schematic diagram of changing lanes;

Figure 8 is a schematic diagram of crossings;

Fig. 9 is a schematic diagram of performing a U-Turn turn;

Fig. 10 is a hardware structural diagram of the vehicle-mounted recording device of the present invention.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be described in further detail below in conjunction with specific embodiments and with reference to the accompanying drawings.

The cognitive ability test system of the unmanned vehicle of the present invention is composed of a complete set of software and hardware platforms, which obtains its cognitive information through a wired or wireless connection with the unmanned vehicle, thereby realizing the test of the cognitive ability of the unmanned vehicle test.

Figure 1 is an overall block diagram of the cognitive ability test system for unmanned vehicles.

Since the tested unmanned vehicles should have basic autonomous driving capabilities on urban roads, intercity roads and rural roads, the present invention starts from the four elements of the road system (people, vehicles, roads and traffic signals), and assumes that Outside the road, pedestrians, objects, street trees and buildings that do not affect the driving of the unmanned vehicle are not within the scope of the present invention.

On the basis of the above assumptions and analysis, the test items of the present invention include basic cognitive ability test items, advanced cognitive ability test items and comprehensive cognitive ability test items.

1) Basic Cognitive Ability Test Items

The cognitive ability of traffic signals, driving lights and whistles (summarized as "visual and auditory information" below) is the basic cognitive ability that unmanned vehicles must have when driving on the road. Therefore, this project mainly tests the recognition ability of unmanned vehicles. Ability to understand traffic signals and driving lights and whistles. This part of the test content is a static test, which is divided into two forms: B-test and I-test according to the degree of difficulty.

B-Test means that the tested vehicle is in a static state, recognizes standard audiovisual information, and requires the unmanned vehicle to output images, texts, symbols or sounds representing the corresponding audiovisual information. The main test index is the correct identification of audiovisual information. number and the time required for recognition. Standard audiovisual information includes the following:

Traffic signals include traffic signs, traffic markings, traffic lights and traffic police commands, see national standards GB5768-2009 and GB14887-2003.

Traffic signs are divided into two categories: main signs and auxiliary signs. The national standard GB5768-2009 "Road Traffic Signs and Markings" is applicable to highways, urban roads and places that are within the jurisdiction of the unit but allow social motor vehicles to pass, including squares, public parking lots and other places used for public traffic, etc. Traffic signs and markings set on the road. The traffic signs and markings set up in places where other motor vehicles pass, parking lots, etc. can be implemented as a reference. The main signs include: ①Warning signs (54 types); ②Prohibition signs (48 types); ③Instruction signs (39 types); ④Guide signs (87 types); ⑤Tourist area signs (17 types); (The signs used in the work area are warning signs, prohibition signs, instruction signs and road signs); ⑦Notice signs (8 types). Prohibition signs and instruction signs are signs that road users must obey; other signs provide information only, such as guide signs, tourist area signs. Auxiliary signs (19 types) are signs attached to the main sign to provide auxiliary explanations.

Traffic markings are traffic facilities composed of various lines, arrows, characters, patterns, facade marks, physical marks, raised road signs and contour marks drawn or installed on the road. Traffic markings are divided into the following three categories according to their functions: 1) indicating markings (35 types); 2) prohibiting markings (21 types); 3) warning markings (9 types).

Traffic signal lights can be divided into motor vehicle signal lights, non-motor vehicle signal lights, pedestrian crossing signal lights, lane signal lights, direction indicator lights, flashing warning lights, and crossing signal lights according to their functions. Some road traffic lights may be equipped with countdown digital display. The cognition of unmanned vehicles to traffic lights may be affected by the arrangement order and installation position of the signal lights. For information on the arrangement order and installation positions of the traffic lights used in the present invention, please refer to the national standard GB14886-2006 "Road Traffic Signal Light Setting and Installation Specification", which is applicable to the installation of signal lights at urban roads and highway level intersections (hereinafter referred to as intersections), urban roads and highway sections (hereinafter referred to as sections), urban roads and highway and railway level intersections (hereinafter referred to as crossings) . The national standard GB14887-2003 "Road Traffic Signal Lights" is applicable to signal lights used on roads. According to the shape of the intersection, it can be divided into: cross, oblique, T, Y, dislocation T, dislocation Y, multi-way and roundabout.

Traffic police command includes traffic police gestures (11 types).

Driving lights refer to the lights used by vehicles in the process of participating in traffic. The lights that are often used in driving mainly include turn signals, headlights, rear position lights, position marker lights, brake lights, etc. Whistle refers to the sound of horns or sirens on special vehicles when vehicles participate in traffic. The special vehicles mentioned here mainly include police cars, fire trucks, ambulances and engineering emergency vehicles.

I-test means that the tested vehicle is in a static state, under the following three conditions: ① cognition of non-standard audio-visual information (standard audio-visual information is affected by shadows, defacement, occlusion, tilting, blurring - wind blowing The impact of fading, light, deformation and other interference factors after sun exposure); ② The extraction and recognition of the corresponding audiovisual information target under the condition of feature binding; ③ The required audiovisual information target under the multi-information condition Extraction and cognition require unmanned vehicles to output images, text, symbols or sounds representing corresponding audiovisual information. The main test indicators of the I test are the number of correctly recognized audiovisual information and the time required for recognition. The schematic diagram of the test scheme is shown in Figure 2.

2) Advanced Cognitive Ability Test

The content of this test is a dynamic test. It is based on the basic cognitive ability test, and the unmanned vehicle and the tested audio-visual information are in a relative motion state to test the impact of the unmanned vehicle on the traffic environment in a certain traffic scene. and Traffic Behavior Cognitive Ability, referred to as the M-test.

The content of this part of the test is to test one or several traffic scenes and traffic phenomena in a relatively short distance. target, speed, position, time, behavior, etc.).

M-tests can be specifically divided into two categories: visual and auditory information is moving while the driverless vehicle is not moving and audiovisual information is not moving while the driverless vehicle is moving. The movement of audiovisual information here mainly refers to the situation that the audiovisual information source (or target source) changes greatly in the far and near positions (longitudinal), while the left and right positions (horizontal) change relatively small (this is more consistent with the actual road conditions, the vehicle distance The source of audiovisual information is always from far to near or from near to far, that is, the longitudinal change is relatively large, but limited by the width of the road and the setting position, the lateral change is relatively small); in driving condition. The schematic diagram of the M test scheme is shown in Figure 3.

The traffic scenes here include intersections, roundabouts, urban expressway ramps, expressway ramps, parking areas, tunnels, overpasses, viaducts, sharp turning roads, uphill roads, downhill roads, roads with lateral inclination angles, etc.; specific traffic phenomena include Turning, changing lanes, overtaking, meeting, reversing, U-turn, merging, parking, crossing traffic, overpass traffic, railway crossing traffic, overpass and railway bridge bottom traffic, emergency braking, pedestrian avoidance, cargo scattered, road Standing obstacles, etc.

3) Comprehensive Cognitive Ability Test

The content of this test is a dynamic test. On the basis of basic cognitive ability and advanced cognitive ability test, it mainly tests the comprehensive cognitive ability of unmanned vehicles to traffic scenes and the ability of autonomous driving and intelligent navigation, referred to as S-test .

The S-test is relatively complex, mainly divided into urban road environment, intercity road environment, rural road environment and road environment combining the three. Within a certain long distance, unmanned vehicles need to identify obstacles, traffic signals, driving Lights, whistles, traffic infrastructure and traffic phenomena, etc., and complete autonomous driving and intelligent navigation according to the requirements. target, speed, position, time, behavior, etc.). For example: in normal traffic flow such as intercity expressways and urban roads, unmanned vehicles are required to be able to communicate with manned vehicles and unmanned vehicles based on audiovisual information (traffic signals, driving lights and whistles) ) for multi-vehicle interaction and cooperative driving of unmanned vehicles, so that multiple unmanned vehicles can be mixed in the normal intercity traffic flow; in a large urban area and under complex traffic flow conditions, unmanned vehicles are required to be able to autonomously Driving, including the detection and recognition of basic traffic signs under the premise of complying with traffic laws and in a controllable real driving environment, vehicle speed control in different traffic environments, lane keeping, dynamic overtaking, static and dynamic obstacle avoidance, and traffic control in specified areas Autonomous parking, and autonomous driving in a variety of controllable urban roads (dense roads, urban ring roads, ring roads, etc.).

In order to achieve the purpose of testing the cognitive ability of unmanned vehicles through the above test items, the test scene of the present invention is not a simple combination of scenes and sensing, measurement, and computing equipment, but for static, dynamic and uncertain environments Complex test scenarios for completing test tasks. The present invention constructs a minimum set of traffic environment and traffic phenomenon as follows: traffic scenes include intersections, roundabouts, urban expressway ramps, expressway ramps, parking areas, tunnels, overpasses, viaducts, sharp turning roads, uphill roads, downhill roads, Roads with a lateral inclination angle, etc.; traffic phenomena include turning, changing lanes, overtaking, meeting vehicles, reversing, turning around, merging vehicles, parking, crossing traffic, overpass traffic, crossing traffic, overpass and railway bridge bottom traffic, stop Line parking, emergency braking, pedestrian avoidance, scattered goods, stationary obstacles on the road, etc.

【Description of test scene design】

The cognitive test standard system for unmanned vehicles of the present invention is composed of basic cognitive ability test, advanced cognitive ability test and comprehensive cognitive ability test. The order from low to high is: B-test, I-test, M-test and S-test, the first two are static tests, and the last two are dynamic tests. In order to describe the content of the present invention in terms of specific test scenario design in more detail, the four test contents will be described in detail below.

B-Test and I-Test Scenarios: For static tests, since the unmanned vehicle under test is in a static state, the corresponding test scenarios are relatively simple, and a relatively wide one can be selected arbitrarily according to the needs of the actual test content. The space is enough, or it can be determined according to the test scene selected by the dynamic test.

M-test and S-test scenarios: For dynamic tests, since the tested unmanned vehicle is in a driving state, on the one hand, it is in consideration of safety; on the other hand, technological progress is a gradual process. Before the test site, it is recommended to select a relatively closed real road environment (such as a driving school test site) as much as possible, or an urban road, an intercity road or a rural road environment that has not been opened to traffic or has a very small traffic flow. It is more appropriate, and then on this basis, select and design the corresponding traffic environment that meets the needs of M-test and S-test according to the actual test content. The M-test and S-test are illustrated below.

M-test example:

At an intersection and its surrounding roads as shown in Figure 4, unmanned vehicles must recognize the information of roadside traffic signs while driving along the road, and enter the corresponding lane according to the instructions of the traffic signs bound to the number , and follow the instructions of the traffic lights, obey the traffic rules, and pass the intersection smoothly.

Note: The cognitive content of the basic cognitive ability test assessed in this example includes: traffic signs, feature binding, traffic lights, traffic markings and intersections; the content of the advanced cognitive ability test assessed: Keep lanes, change lanes, stop at stop lines, and pass intersections.

As far as this example is concerned, it is also possible to set whether the sign is affected by interference factors, whether there are obstacles on the road, whether to add pedestrian crossing markings at the intersection, whether there are pedestrians passing, whether there are other manned or unmanned vehicles on the road Vehicles, the density of vehicles, whether there will be multiple information in the field of vision of unmanned vehicles, etc., to design and increase the content and difficulty of the test.

S-test example:

During the entire test process, unmanned vehicles are required to drive according to the traffic signs on the road, and must perform the following operations in the process of completing the race: lane keeping, stopping at the stop line, changing lanes, passing intersections, U-turn, etc. . The specific test requirements are as follows:

① Lane keeping driving: Unmanned vehicles are driving on structured roads. Except for turning, overtaking, obstacle avoidance, U-Turn and other actions, unmanned vehicles cannot leave the lane and should keep driving in the lane. The turning curve at the intersection should be a smooth curve connecting the two road sections. Figure 5 is a schematic diagram of the driving trajectory of the unmanned vehicle keeping the lane when turning left through the intersection.

②Stopping at the stop line: Unmanned vehicles are required to be able to identify the stop line and execute the stop. When an unmanned vehicle encounters a stop sign or stop line during driving, it should stop within one meter of the stop line (as shown in Figure 6). A driverless vehicle cannot stop across a stop line at an intersection.

③ Lane change: When a moving object is found in front, the unmanned vehicle is required to change lanes within a certain distance from the moving object in front (as shown in Figure 7).

④Through the intersection: When passing the intersection, follow the instructions of the traffic lights; do not enter the intersection when there is a traffic jam at the intersection; when turning left to pass the intersection, you must turn left at the center of the intersection (as shown in Figure 8).

⑤ U-Turn: Unmanned vehicles should comply with traffic laws and regulations when performing U-Turn, and must not hinder the normal driving of other vehicles and pedestrians (as shown in Figure 9).

[Communication interface and communication protocol design description] In order to provide unmanned vehicles with test tasks that they can recognize, it is necessary to input task files that it can recognize to unmanned vehicles. The communication interface and communication protocol of the present invention stipulate transmission equipment Method and format for transmitting test tasks to autonomous vehicles. At the same time, in order to accurately evaluate the cognitive ability of unmanned vehicles, quantify the evaluation indicators, achieve well-documented, and ensure the "fairness, justice, and openness" of the test, the present invention must have an effective recording device to collect and Store the state of the unmanned vehicle and the cognitive results of the target. Based on the above considerations, the present invention designs a vehicle-mounted recording device, which has the functions of a transmission device and a recording device: input cognitive ability test tasks to the unmanned vehicle; receive and store data output by the unmanned vehicle cognitive ability test.

The test task input program and the test data output program of the driverless vehicle cognitive ability test are two independent programs. After the unmanned vehicle receives the test task through the test task input program, it should re-establish the serial port connection, and open the corresponding program to output the test data to the on-board recording device. Considering the ease of use of the program, the test task input and test data output of the driverless vehicle cognitive ability test can adopt a unified interface, parameters and data transmission format, but there are differences in the specific format of the data content.

In order to ensure the reliability of the communication between the unmanned vehicle and the on-board recording equipment, considering the size of the data required for communication, the variety of vehicle equipment, and the different interfaces, the present invention adopts the RS-232 interface with strong versatility for information transmission. The unmanned vehicle needs to have a DB9 serial port male connector placed in the vehicle for easy insertion, and reserve the necessary space for the on-board recording equipment. The on-board recording device uses its own power supply and provides a DB9 serial port female connector. The hardware structure diagram is shown in Figure 10.

Referring to FIG. 10 , the on-board recording device further includes a storage module, a backup battery, a test task output module and a test data input module. In order to ensure the working time of the on-board recording equipment, a backup battery module is set up, which can automatically switch to the backup battery when the power of the currently used battery is lower than a certain value. After the on-board recording device starts working, the task stored in the device is first sent to the unmanned vehicle through the test task output module, and then the test data input module takes over, waiting for the test result sent back by the unmanned vehicle and saving the result into the storage module.

In order to communicate with an autonomous vehicle, an efficient communication protocol must be established. In the entire communication system, the on-board recording device is responsible for transmitting handshake packets and data packets as the master device; while the unmanned vehicle as the slave device only replies to handshake packets and ACK packets. The interface of the communication interface device preferably adopts RS-232, and is connected in a wired manner.

Communications between on-board recording devices and unmanned vehicles need to meet the following requirements:

Communication parameters: Baud rate 9600bps, 8 data bits, 1 stop bit, no parity.

Communication format:

package type Data length data content

Packet type: Packet types include handshake packets, ACK packets and data packets.

Handshake packet format:

Request direction:

0: Indicates that the unmanned vehicle requests to send data to the on-board recording device;

1: Indicates that the on-board recording device requests to send data to the driverless vehicle.

Sending and receiving status:

0: Indicates that it is not ready to receive or transmit data;

1: Indicates that it is ready to receive or transmit data.

ACK packet format:

package type Data length data content ACK packet 3 1 byte ACK status

0 or 1

ACK status:

0: Indicates that the data received by the receiving end is incorrect;

1: Indicates that the correct receiving end has received the data.

Note: If the sending end (vehicle recording device) does not receive the ACK packet from the receiving end (unmanned vehicle) within 100ms after sending the data, the data will be resent.

Packet format:

Data content format:

Note: Before a file is transmitted, the on-board recording device will calculate the file frame number of the current data packet according to the size of the file, and add it to the file frame number of each data packet in reverse order. When the unmanned vehicle receives a data packet, it should check the file frame number in the data content, and splice the files transmitted in subpackages; when the unmanned vehicle receives duplicate file frame numbers, it needs to discard the duplicate frame numbers The data; when the unmanned vehicle receives the data packet with file frame number 0, the file transmission ends.

Examples of input tasks for the cognitive ability test of unmanned vehicles:

① Connect the serial port, the referee sends a handshake packet to the unmanned vehicle through the on-board recording device, sets the sending and receiving status bit to 1, and the request direction to 1, requesting to transmit the task file;

② If the unmanned vehicle is not ready to receive data, it will reply the handshake packet, set the sending and receiving status bit to 0, and the request direction to 1”;

③The on-board recording device checks whether the handshake packet is received within 100ms. If the handshake packet is not received within 100ms, it will resend the handshake packet until it receives a handshake packet from the unmanned vehicle to indicate that it is ready;

④ After receiving the handshake packet, the on-board recording device starts to send the first data packet;

⑤ After the unmanned vehicle receives the data packet, it checks the check code. If the check code is incorrect, it replies with ACK of 0, indicating a request for retransmission;

⑥If the check code is correct, the unmanned vehicle first checks the file frame number in the data content. When the file frame number is 0, it means that the file transmission is over, and then the unmanned vehicle replies with an ACK of 1, indicating confirmation of receipt. Otherwise, dump the current packet data, clear the buffer to receive the next data packet, and reply ACK as 1;

⑦ If the on-board recording device receives ACK is 0, then resend the data packet;

⑧If the on-board recording device does not receive any ACK packet within 100ms, then resend the data packet until the ACK returned by the unmanned vehicle is 1;

⑨When the on-board recording device receives the ACK from the unmanned vehicle as 1, it sends the next data packet;

⑩The unmanned vehicle continues to check the file frame number of the data content. When the file frame number is 0, the file transfer ends.

The output data interface and communication protocol of the unmanned vehicle cognitive ability test. This agreement specifies the method and format of the unmanned vehicle's transmission of test data to the on-board recording device. As the master device, the unmanned vehicle is responsible for transmitting handshake packets and data packets, and the on-board recording device, as a slave device, only replies to handshake packets and ACK packets. In order to record the location information of the unmanned vehicle throughout the process, the unmanned vehicle needs to regularly send test data to the on-board recording device; at the same time, in order to check the results of the unmanned vehicle's detection target at the first time, the unmanned vehicle needs to immediately send the detection target Data, that is, the unmanned vehicle needs to send the data content in two formats of regular data format and real-time data format to the on-board recording device respectively.

The vehicle-mounted recording device performs time statistics on received timing data format data packets according to its own clock. If the unmanned vehicle sends data within a time much shorter than the specified interval, the vehicle-mounted recording device will no longer record. Therefore, unmanned vehicles should avoid sending a large number of data packets in a short period of time. The on-board recording device will perform time and content statistics on the received instant data format packets, see instant data format for details. The on-vehicle recording device does not repeatedly record any data content with repeated data packet numbers.

Communication interface: RS-232, wired connection.

Communication parameters: Baud rate 9600bps, 8 data bits, 1 stop bit, no parity.

Communication format:

package type Data length data content

Packet type: Packet types include handshake packets, ACK packets and data packets.

Handshake packet format:

Request direction:

0: Indicates that the unmanned vehicle requests to send data to the on-board recording device;

1: Indicates that the on-board recording device requests to send data to the driverless vehicle.

Sending and receiving status:

0: Indicates that it is not ready to receive or transmit data;

1: Indicates that it is ready to receive or transmit data.

ACK packet format:

ACK status:

0: Indicates that the data received by the receiving end is incorrect;

1: Indicates that the correct receiving end has received the data.

Note: If the sending end (unmanned vehicle) does not receive the ACK packet from the receiving end (vehicle recording device) within 100ms after sending the data, the data will be resent.

Packet format:

package type Data length data content data pack 2 fixed byte Test Data

Data content format: The data content format is divided into timing data and real-time data format, which are distinguished by data packet number.

Note: The keyword length and keyword expression method in the data content format can be determined according to specific needs.

Timing data format: The content of the timing data format is the location and time information sent by the unmanned vehicle at regular intervals.

Instant data format: The content of the instant data format is the real-time information of the target detection sent by the unmanned vehicle.

Data Packet Number: It is the unique number for each data pack. In the timing data format, start with the ASCII code of T; in the immediate data format, start with the ASCII code of R.

Vehicle Number: The unique designation assigned to the team prior to the start of the race.

Time: Time information when data is sent.

Location: Information about the location of the driverless vehicle at the time the data was sent.

Target information: The target information detected by the unmanned vehicle. The target information format is as follows:

Object Code: Includes type codes and their corresponding numbers.

Note: In order to avoid communication congestion and storage overflow of the vehicle-mounted recording device, the vehicle-mounted recording device will only record data packets in the instant data format with repeated data packet numbers or target codes within a certain period of time (such as 1 second). The information is subject to, and the part that is sent repeatedly will not be received or stored, that is, after the unmanned vehicle detects the target to be tested, it should only send the data about the target once.

Target status: Indicates the speed information of the measured target.

Example of data output for driverless vehicle cognitive ability test:

① Connect the serial port, the unmanned vehicle sends a handshake packet to the on-board recording device, set the sending and receiving status bit to 1, request the direction to 0, and request to transmit test data.

② If the on-board recording device is not ready to receive data, it will reply the handshake packet, set the sending and receiving status bit to 0, and the request direction to 0".

③The unmanned vehicle checks whether the handshake packet is received within 100ms, and resends the handshake packet if it has not been received within 100ms. Until the on-board recording device replies with a handshake packet indicating that it is ready.

④ After the unmanned vehicle receives the handshake packet, it starts to send the first data packet.

⑤ After the vehicle-mounted recording device receives the data packet, if the check code is incorrect, it will reply ACK with 0, indicating a request for retransmission.

⑥If the verification code is correct, the on-board recording device will reply with ACK of 1, which means confirmation of receipt.

⑦ If the ACK received by the unmanned vehicle is 0, the data packet will be retransmitted.

⑧ If the unmanned vehicle does not receive any ACK packet within 100ms, it will resend the data packet. Until the ACK returned by the on-board recording device is 1.

⑨The unmanned vehicle sends the next data packet after receiving the ACK returned by the on-board recording device as 1.

The specific embodiments described above have further described the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above descriptions are only specific embodiments of the present invention, and are not intended to limit the present invention. Within the spirit and principles of the present invention, any modifications, equivalent replacements, improvements, etc., shall be included in the protection scope of the present invention.

Claims (24)

1. An unmanned vehicle cognitive ability testing system, the system includes an unmanned vehicle and a vehicle-mounted recording device, both of which are wired or wirelessly connected through a communication interface, characterized in that, The on-board recording device is used to input cognitive ability test tasks to the unmanned vehicle, and receive and store the output data of the unmanned vehicle cognitive ability test; After the unmanned vehicle receives the test task through the test task input program, re-establish the serial port connection, open the corresponding program to output the test data to the on-board recording device, the test task input and test data output of the unmanned vehicle cognitive ability test adopt a unified The interface, parameters and data transmission format, but the specific format of the data content is different, Among them, the test tasks include basic cognitive ability test items, advanced cognitive ability test items and comprehensive cognitive ability test items. 2. The system according to claim 1, characterized in that the basic cognitive ability test item tests the driverless vehicle's ability to recognize traffic signals and driving lights and whistles. 3. The system according to claim 2, wherein the basic cognitive ability test includes B-test and I-test, wherein B-test means that the vehicle under test is in a stationary state, recognizes standard audio-visual information, and requires Unmanned vehicles output images, texts, symbols or sounds representing corresponding audiovisual information, and the test indicators are the number of correctly identified audiovisual information and the time required for identification; I-test means that the tested vehicle is in a static state, The following three situations: ① cognition of non-standard audiovisual information; ② extraction and recognition of corresponding audiovisual information targets under the condition of feature binding; ③ required audiovisual information under the condition of multiple information The extraction and cognition of the target requires the unmanned vehicle to output images, texts, symbols or sounds representing the corresponding audio-visual information. The I test is the number of correctly identified audio-visual information and the time required for identification. 4. The system according to claim 3, characterized in that the advanced cognitive ability test belongs to the dynamic test, which is based on the basic cognitive ability test, the unmanned vehicle and the tested audiovisual information are in a relative motion state Next, test the cognition ability of unmanned vehicles to the traffic environment and traffic behavior in certain traffic scenarios. 5. The system according to claim 4, characterized in that the advanced cognitive ability test includes two types: one is that the driverless vehicle does not move while seeing and hearing information, and the other is that the unmanned vehicle does not move while seeing and hearing information. Unmanned vehicles move. 6. The system according to any one of claims 1-5, wherein the comprehensive cognitive ability test is a dynamic test, and on the basis of the basic cognitive ability and advanced cognitive ability tests, the test of unmanned vehicles is The comprehensive cognitive ability of traffic scenes and the ability of autonomous driving and intelligent navigation. 7. The system according to claim 6, wherein the comprehensive cognitive ability test is divided into urban road environment, intercity road environment, country road environment and road environment combining the three, within a certain long distance, Unmanned vehicles need to identify obstacles, traffic signals, driving lights, sirens, traffic infrastructure and traffic phenomena, and complete autonomous driving and intelligent navigation as required. time, completion of assigned tasks. 8. The system according to claim 1, wherein when the vehicle-mounted recording device transmits data to the unmanned vehicle, the vehicle-mounted recording device is responsible for transmitting handshake packets and data packets as a master device; Reply handshake packet and ACK packet. 9. The system according to claim 8, characterized in that, before a file starts to be transmitted, the vehicle-mounted recording device will calculate the file frame number of the current data packet according to the size of the file, and will add it to each data packet in reverse order In the file frame number, when the unmanned vehicle receives a data packet, it should check the file frame number in the data content, and splice the files transmitted in sub-packets; when the unmanned vehicle receives duplicate file frame numbers , the data with repeated frame numbers needs to be discarded; when the unmanned vehicle receives the data packet with file frame number 0, the file transfer ends. 10. The system according to claim 1, wherein when the unmanned vehicle transmits test data to the on-board recording device, the unmanned vehicle is responsible for transmitting handshake packets and data packets as the master device, and the on-board recording device is used as the slave device Only the handshake packet and ACK packet are replied, and the unmanned vehicle regularly sends test data to the on-board recording device. 11. The system according to claim 10, characterized in that the unmanned vehicle immediately sends the data of the detection target, that is, the unmanned vehicle needs to send data in two formats of timing data format and real-time data format to the on-board recording device respectively content. 12. The system according to claim 11, wherein the vehicle-mounted recording device performs time statistics on the received timing data format data packets according to its own clock, and if the unmanned vehicle sends data in a time far less than the specified interval, Then the vehicle-mounted recording device will no longer record, and the vehicle-mounted recording device will perform time and content statistics on the received instant data format data packets, and the vehicle-mounted recording device will not repeatedly record any data content with repeated data packet numbers. 13. A method for testing the cognitive ability of an unmanned vehicle, characterized in that the method comprises the steps of: The on-board recording device is used to input cognitive ability test tasks to the unmanned vehicle, and receive and store the output data of the unmanned vehicle cognitive ability test; After the unmanned vehicle receives the test task through the test task input program, re-establish the serial port connection with the on-board recording device, open the corresponding program to output the test data to the on-board recording device, the test task input and The test data output adopts a unified interface, parameters and data transmission format, but the specific format of the data content is different. Among them, the test tasks include basic cognitive ability test items, advanced cognitive ability test items and comprehensive cognitive ability test items. 14. The method according to claim 13, characterized in that the basic cognitive ability test item tests the driverless vehicle's ability to recognize traffic signals, driving lights and whistles. 15. The method according to claim 13, wherein the basic cognitive ability test includes B-test and I-test, wherein B-test means that the vehicle under test is in a stationary state, recognizes standard audiovisual information, and requires Unmanned vehicles output images, texts, symbols or sounds representing corresponding audiovisual information, and the test indicators are the number of correctly identified audiovisual information and the time required for identification; I-test means that the tested vehicle is in a static state, The following three situations: ① cognition of non-standard audiovisual information; ② extraction and recognition of corresponding audiovisual information targets under the condition of feature binding; ③ required audiovisual information under the condition of multiple information The extraction and cognition of the target requires the unmanned vehicle to output images, texts, symbols or sounds representing the corresponding audio-visual information. The I test is the number of correctly identified audio-visual information and the time required for identification. 16. The method according to claim 14, wherein the advanced cognitive ability test is a dynamic test, which is based on the basic cognitive ability test, and the unmanned vehicle and the tested audiovisual information are in a relative motion state Next, test the cognition ability of unmanned vehicles to the traffic environment and traffic behavior in certain traffic scenarios. 17. The method according to claim 15, characterized in that the advanced cognitive ability test includes two types: one is to see and hear information while the unmanned vehicle is still, and the other is to see and hear information without moving Unmanned vehicles move. 18. The method according to any one of claims 13-17, characterized in that the comprehensive cognitive ability test is a dynamic test, and on the basis of the basic cognitive ability and advanced cognitive ability tests, the unmanned vehicle is tested against The comprehensive cognitive ability of traffic scenes and the ability of autonomous driving and intelligent navigation. 19. The method according to claim 13, wherein the comprehensive cognitive ability test is divided into road environments facing urban road environments, intercity road environments, rural road environments and the combination of the three, within a certain long distance, Unmanned vehicles need to identify obstacles, traffic signals, driving lights, sirens, traffic infrastructure and traffic phenomena, and complete autonomous driving and intelligent navigation as required. time, completion of assigned tasks. 20. The method according to claim 13, characterized in that, when the vehicle-mounted recording device transmits data to the unmanned vehicle, the vehicle-mounted recording device is responsible for transmitting handshake packets and data packets as a master device; Reply handshake packet and ACK packet. 21. The method according to claim 20, characterized in that, before a file starts to be transmitted, the vehicle-mounted recording device will calculate the file frame number of the current data packet according to the size of the file, and add it to each data packet in reverse order In the file frame number, when the unmanned vehicle receives a data packet, it should check the file frame number in the data content, and splice the files transmitted in sub-packets; when the unmanned vehicle receives duplicate file frame numbers , the data with repeated frame numbers needs to be discarded; when the unmanned vehicle receives the data packet with file frame number 0, the file transfer ends. 22. The method according to claim 13, characterized in that, when the unmanned vehicle transmits test data to the on-board recording device, the unmanned vehicle is responsible for transmitting handshake packets and data packets as the master device, and the on-board recording device is used as the slave device Only the handshake packet and ACK packet are replied, and the unmanned vehicle regularly sends test data to the on-board recording device. 23. The method according to claim 22, characterized in that the unmanned vehicle immediately sends the data of the detection target, that is, the unmanned vehicle needs to send the data in two formats of timing data format and real-time data format to the on-board recording device respectively content. 24. The method according to claim 23, wherein the vehicle-mounted recording device performs time statistics on the received timing data format data packets according to its own clock, and if the unmanned vehicle sends data within a time much shorter than the specified interval, Then the vehicle-mounted recording device will no longer record, and the vehicle-mounted recording device will perform time and content statistics on the received instant data format data packets, and the vehicle-mounted recording device will not repeatedly record any data content with repeated data packet numbers.