JP7537375B2 - BEHAVIOR PREDICTION SYSTEM, BEHAVIOR PREDICTION DEVICE, BEHAVIOR PREDICTION METHOD, BEHAVIOR PREDICTION PROGRAM - Google Patents

Description

This disclosure relates to technology for predicting the behavior of road users.

Patent Document 1 discloses a technology for predicting the future position of a moving object. With this technology, the future position of a moving object is predicted based on environmental information related to the driving environment. Patent Document 1 discloses, as environmental information, map information such as roads, driving lanes, parking lots, and the radius of curvature of curves, topographical information such as altitude and road gradient, road surface conditions such as puddles, snow accumulation, and fallen objects, and traffic information such as traffic light status.

JP 2018-55141 A

In recent years, there has been an increasing need to predict the behavior of road users more reliably. There is room for improvement in the reliability of the prediction of the behavior of moving objects in Patent Document 1.

An object of the present disclosure is to provide a behavior prediction system with improved reliability. Another object of the present disclosure is to provide a behavior prediction device with improved reliability. Yet another object of the present disclosure is to provide a behavior prediction method with improved reliability. Yet another object of the present disclosure is to provide a behavior prediction program with improved reliability.

The technical means of the present disclosure for solving the problems will be explained below. Note that the claims and the reference characters in parentheses in this section indicate the corresponding relationship with the specific means described in the embodiments described in detail later, and do not limit the technical scope of the present disclosure.

A first aspect of the present disclosure is a behavior prediction system for predicting a behavior of a road user, the system having a processor (102),
The processor
Obtaining road user identification information for identifying a road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
Preparing correlation information between road users and structures as destinations of the road users;
predicting road user behavior based on road user identification information, structure identification information, and correlation information;
is configured to run
Preparing the correlation information includes preparing map information, for each type of road user, in which the structure user distribution estimated for each structure is at least associated as structure identification information;
Predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and map information for each type of road user .

A second aspect of the present disclosure is a behavior prediction device for predicting a behavior of a road user, the behavior prediction device having a processor (102),
The processor
Obtaining road user identification information for identifying a road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
Preparing correlation information between road users and structures as destinations of the road users;
predicting road user behavior based on road user identification information, structure identification information, and correlation information;
is configured to run
Preparing the correlation information includes preparing map information, for each type of road user, in which the structure user distribution estimated for each structure is at least associated as structure identification information;
Predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and map information for each type of road user .

A third aspect of the present disclosure provides a behavior prediction method executed by a processor (102) for predicting a behavior of a road user, the method comprising:
Obtaining road user identification information for identifying a road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
Preparing correlation information between road users and structures as destinations of the road users;
predicting road user behavior based on road user identification information, structure identification information, and correlation information;
Including,
Preparing the correlation information includes preparing map information, for each type of road user, in which the structure user distribution estimated for each structure is at least associated as structure identification information;
Predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and map information for each type of road user .

A fourth aspect of the present disclosure is a behavior prediction program stored in a storage medium (101) for predicting a behavior of a road user, the behavior prediction program including instructions to be executed by a processor (102), the program comprising:
The command is,
Obtaining road user identification information that identifies a road user;
Obtaining structure identification information that identifies a structure that the road user can enter and exit;
preparing correlation information between road users and structures as destinations of the road users;
Predicting the behavior of road users based on road user identification information, structure identification information, and correlation information;
Including,
Preparing the correlation information includes preparing map information, in which the structure user distribution estimated for each structure is at least associated as structure identification information, for each type of road user;
Predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and map information for each type of road user .
According to the first to fourth aspects, the behavior of the road user is predicted based on the road user identification information, the structure identification information, and the correlation information between the road user and the structure. According to this, the correlation between the road user and the structure as the destination is taken into consideration in predicting the behavior of the road user. Therefore, the behavior of the road user who has the structure as the destination can be included in the prediction result. Therefore, the reliability of the behavior prediction can be improved.

FIG. 1 is a block diagram showing an overall configuration of a first embodiment. FIG. 1 is a diagram illustrating an application environment of a first embodiment. 1 is a block diagram showing a functional configuration of a behavior prediction system according to a first embodiment. FIG. 4 is a diagram showing an example of structure mapping information in the first embodiment. FIG. 13 is a diagram illustrating an example of a data structure of correlation information. 3 is a flowchart illustrating a behavior prediction method according to the first embodiment. FIG. 11 is a diagram showing an example of structure mapping information in the second embodiment.

Below, multiple embodiments of the present disclosure are described with reference to the drawings. Note that in each embodiment, corresponding components are given the same reference numerals, and duplicated descriptions may be omitted. Furthermore, when only a portion of the configuration is described in each embodiment, the configuration of the other embodiment described above can be applied to the other portions of the configuration. Furthermore, in addition to the combinations of configurations explicitly stated in the description of each embodiment, configurations of multiple embodiments can be partially combined together even if not explicitly stated, as long as there is no particular problem with the combination.

An embodiment of the present disclosure is described below with reference to the drawings.

First Embodiment
The behavior prediction system 100 of the first embodiment shown in FIG. 1 predicts future behavior of road users existing in the vicinity of the host vehicle A shown in FIG. 2. The road users are moving objects that use the road. The roads here may include roadways, sidewalks, and other adjacent spaces. The moving objects here include the target vehicle B and pedestrians. From a perspective centered on the host vehicle A, the host vehicle A can also be said to be an ego-vehicle. From a perspective centered on the host vehicle A, the target vehicle B can also be said to be another vehicle, or another road user. The target vehicle B includes four-wheeled automobiles and two-wheeled vehicles. The target vehicle B may be an unmanned vehicle without a driver on board. The host vehicle A can communicate with the target vehicle B and the center C through an external network NW.

In the host vehicle A, an autonomous driving mode is provided, which is classified into levels according to the degree of manual intervention by the occupant in the driving task. The autonomous driving mode may be realized by autonomous driving control, such as conditional driving automation, high driving automation, or full driving automation, in which the system performs all driving tasks when activated. The autonomous driving mode may also be realized by advanced driving assistance control, such as driving assistance or partial driving automation, in which the occupant performs some or all driving tasks. The autonomous driving mode may be realized by either one of the autonomous driving control and the advanced driving assistance control, or a combination of them, or by switching between them.

The host vehicle A is equipped with a sensor system 10, a map database (hereinafter, "DB") 20, and a communication system 30, as shown in FIG. 3. The sensor system 10 acquires sensor information that can be used by the behavior prediction system 100 by detecting the external and internal environments of the host vehicle A. To this end, the sensor system 10 is configured to include an external sensor 11 and an internal sensor 12.

The external sensor 11 acquires external information that can be used by the behavior prediction system 100 from the external world that is the surrounding environment of the host vehicle A. The external sensor 11 may acquire the external information by detecting targets that exist in the external world of the host vehicle A. The target detection type external sensor 11 is at least one of the following types: a camera, LiDAR (Light Detection and Ranging / Laser Imaging Detection and Ranging), radar, sonar, etc.

The internal sensor 12 acquires internal information that can be used by the behavior prediction system 100 from the internal world that is the internal environment of the host vehicle A. The internal sensor 12 may acquire internal information by detecting a specific physical quantity of motion in the internal world of the host vehicle A. The internal sensor 12 of the physical quantity detection type is at least one of a driving speed sensor, an acceleration sensor, a gyro sensor, etc. The internal sensor 12 may acquire internal information by detecting a specific state of an occupant in the internal world of the host vehicle A. The internal sensor 12 of the occupant detection type is at least one of a driver status monitor (registered trademark), a biological sensor, a seating sensor, an actuator sensor, an in-vehicle equipment sensor, etc.

The map DB 20 stores map information that can be used by the behavior prediction system 100. The map DB 20 includes at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium. The map DB 20 may be a database of a locator that estimates the self-state quantity including the self-position of the host vehicle A. The map DB 20 may be a database of a navigation unit that navigates the driving route of the host vehicle A. The map DB 20 may be configured by combining multiple types of these databases.

The map DB 20 acquires and stores the latest map information, for example, by communicating with an external center. Here, the map information is digitized in two or three dimensions as information representing the driving environment of the host vehicle A. In particular, as the three-dimensional map data, digital data of a high-precision map is preferably used. The map information may include road information representing at least one of the following: the position, shape, and road surface condition of the road itself. The map information may include marking information representing at least one of the following: the position and shape of signs and dividing lines attached to the road. The map information may include traffic light information representing at least one of the following: the position and shape of traffic lights. The map information may include structure identification information, which will be described later, for structures that the user can enter and exit.

The communication system 30 acquires communication information usable by the behavior prediction system 100 by wireless communication. The communication system 30 may receive a positioning signal from a satellite of the Global Navigation Satellite System (GNSS) present in the external world of the host vehicle A. The positioning type communication system 30 is, for example, a GNSS receiver. The communication system 30 may transmit and receive communication signals between the host vehicle A and a V2X system present in the external world of the host vehicle A. The V2X type communication system 30 is, for example, at least one of a Dedicated Short Range Communications (DSRC) communication device and a cellular V2X (C-V2X) communication device. The communication system 30 may transmit and receive communication signals between the host vehicle A and a terminal present in the internal world of the host vehicle A. The terminal communication type communication system 30 is, for example, at least one of a Bluetooth (registered trademark) device, a Wi-Fi (registered trademark) device, and an infrared communication device. The communication system 30 enables the host vehicle A to connect to the external network NW. This allows host vehicle A to communicate with center C via communication system 30.

The center C is equipped with a structure DB 40 shown in FIG. 3. The structure DB 40 is a database that stores characteristic information of structures that can be used by the behavior prediction system 100. The structure DB 40 is configured to include at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium. The structure DB 40 may be a database installed in the host vehicle A. The structure DB 40 may be configured by a combination of databases installed in the host vehicle A and the center C.

The characteristic information stored in the structure DB 40 is information about structures that can be used in predicting the behavior of road users. In particular, the characteristic information includes structure characteristics that can estimate the correlation between road users and structures selected as the destinations of the road users. For example, the characteristic information includes information about users (structure users) of individual structures or types of structures. As a specific example, one piece of characteristic information is information about the number of users by time period, such as day of the week or time of day. As another specific example, one piece of characteristic information is attribute information about the user attributes of each structure. For example, the attribute information may include the distribution of structure users by age group, gender, whether or not they are accompanied, and by type. The types of accompanying persons are, for example, families, couples, friends, etc. In addition, the attribute information may also include the distribution of users by other social attributes. Other social attributes include, for example, categories such as company employees, students, and housewives.

The structure DB 40 may update the characteristic information to the latest information as appropriate. The structure DB 40 may update the characteristic information based on information distributed from other information management centers. The structure DB 40 may also update the characteristic information based on detection information (external environment information) of structures by vehicles on the road, such as the host vehicle A and the target vehicle B.

The behavior prediction system 100 is connected to the sensor system 10, the communication system 30, the map DB 20, the communication system 30, and the structure DB 40 via at least one of, for example, a LAN (Local Area Network) line, a wire harness, an internal bus, and a wireless communication line. The behavior prediction system 100 is configured to include at least one dedicated computer.

The dedicated computer constituting the behavior prediction system 100 may be an integrated ECU (Electronic Control Unit) that integrates the driving control of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be a judgment ECU that judges the driving task in the driving control of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be a monitoring ECU that monitors the driving control of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be an evaluation ECU that evaluates the driving control of the host vehicle A.

The dedicated computer constituting the behavior prediction system 100 may be a navigation ECU that navigates the driving route of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be a locator ECU that estimates the self-state quantity of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be an actuator ECU that controls the driving actuator of the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be an HCU (Human Machine Interface (HMI) Control Unit) that controls the information presentation by the information presentation system in the host vehicle A. The dedicated computer constituting the behavior prediction system 100 may be a computer other than the host vehicle A that constitutes, for example, an external center or a mobile terminal capable of communicating with the host vehicle A.

The dedicated computer constituting the behavior prediction system 100 has at least one memory 101 and one processor 102. The memory 101 is at least one type of non-transitory tangible storage medium, such as a semiconductor memory, a magnetic medium, or an optical medium, that non-temporarily stores computer-readable programs and data. The processor 102 includes at least one type of core, such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a RISC (Reduced Instruction Set Computer)-CPU, a DFP (Data Flow Processor), or a GSP (Graph Streaming Processor).

In the behavior prediction system 100, the processor 102 executes a number of instructions contained in a behavior prediction program stored in the memory 101 in order to predict the behavior of road users. In this way, the behavior prediction system 100 constructs a number of functional blocks for predicting the behavior of road users. The multiple functional blocks constructed in the behavior prediction system 100 include a road user acquisition block 110, a structure acquisition block 120, an input information preparation block 130, a behavior prediction block 140, and a trajectory generation block 150, as shown in FIG. 3.

The road user acquisition block 110 acquires road user identification information that identifies a road user. The road user identification information includes at least one of the road user's type, position, speed, acceleration, etc. The road user acquisition block 110 may acquire the road user identification information from the external sensor 11 in the sensor system 10. Alternatively, the road user acquisition block 110 may acquire the road user identification information from the communication system 30.

The structure acquisition block 120 acquires structure identification information that identifies structures that road users can enter and exit. For example, the structure identification information includes information on the type of structure. The type of structure may be determined, for example, according to the granularity of the structure type information contained in the map information stored in the map DB 20. As an example, the type of structure includes at least one of shopping malls, hospitals, companies, police stations, fire stations, parks, and schools.

The structure identification information also includes information regarding the position of the structure. The information regarding the position includes the coordinates of the structure. The information regarding the position may also include size, orientation, and the like. For example, when the coordinates of the structure are the representative position (central position, etc.) of the structure, the size is obtained. The structure identification information also includes information regarding the entrance/exit of the structure. The information regarding the entrance/exit may include information regarding the type of road users who can enter and exit the entrance/exit. The structure identification information also includes characteristic information.

The structure acquisition block 120 may acquire the above structure identification information, for example, by estimation based on external environment information from the external environment sensor 11 in the sensor system 10. The structure acquisition block 120 may acquire the external environment information itself including the structure identification information. Alternatively, the structure acquisition block 120 may acquire the structure identification information from the map information stored in the map DB 20. The structure acquisition block 120 may acquire the map information itself in which the structure identification information is already stored.

The input information preparation block 130 prepares at least a portion of the information used in behavior prediction, i.e., the input information to be input to the behavior prediction model. The prepared input information includes, for example, structure mapping information and correlation information.

The structure mapping information is map information in which the structure-specific information is expanded. The map information in which the structure-specific information is expanded can also be said to be map information to which the structure-specific information is associated. The map information is, for example, two-dimensional map data acquired from the map DB 20. The map information relates to a specific area around the host vehicle A, such as a specific area centered on the host vehicle A. In the example shown in FIG. 4, the input information preparation block 130 associates at least the type and position of the structure with the map information. In addition, in the example shown in FIG. 4, the input information preparation block 130 associates information regarding the entrances and exits of the structure with the map information.

In addition to the structure-specific information, the input information preparation block 130 may also associate road environment information, traffic information, and the like with the map information.

The input information preparation block 130 may generate structure mapping information by associating structure identification information with map information that does not have structure identification information. Alternatively, the input information preparation block 130 may generate structure mapping information by associating other structure identification information with map information in which at least one type of structure identification information is pre-stored. Alternatively, the input information preparation block 130 may use the map information itself, which is acquired by the structure acquisition block 120 and in which structure identification information is pre-stored, as the structure mapping information.

The correlation information is information indicating the correlation between road users and structures as the destinations of the road users. For example, the correlation information has a data structure including a node indicating the type of road user (road user node), a node indicating the type of structure (structure node), and an edge indicating the correlation between the road user node and the structure node, as shown in FIG. 5. In the example shown in FIG. 5, passenger cars, ambulances, trucks, and police cars are set as road user nodes as types of target vehicle B as a road user. Shopping malls, hospitals, companies, and police stations are set as structure nodes as types of structures that can be the destinations of target vehicle B. Each road user node is linked by an edge to a structure node that is assumed to be a destination corresponding to the road user node.

In addition, weights may be set for the edges. For example, the weight of the edge between a police car and a police station may be set to be greater than the weight of the edge between a passenger car and a police station.

The input information preparation block 130 may prepare the correlation information by reading the above correlation information from the memory 101 in which the information is pre-stored. Alternatively, the input information preparation block 130 may prepare the correlation information by acquiring pre-stored correlation information from the structure DB 40. Alternatively, the input information preparation block 130 may generate at least a part of the correlation information based on outside world information, communication information, map information, etc.

The behavior prediction block 140 predicts the behavior of the road user based on the road user identification information, structure mapping information, and correlation information. Specifically, the behavior prediction block 140 predicts the behavior by inputting each piece of information into a trained model that has been trained to output a predicted result of the road user's behavior in response to the input of the above information. Even if there are multiple road users, the behavior prediction block 140 only needs to obtain a predicted result for each road user by inputting the road user identification information of each road user, as well as common structure mapping information and correlation information.

The trajectory generation block 150 generates a future trajectory for the host vehicle A in the autonomous driving mode of the host vehicle A. The trajectory generation block 150 generates a future trajectory for the host vehicle A based on the prediction result of the behavior prediction block 140. The future trajectory is a planned driving trajectory according to the future behavior, and specifies the driving position of the host vehicle A according to the progress of the host vehicle A. In addition, the future trajectory may specify the speed of the host vehicle A at each driving position. The trajectory generation block 150 may generate a future trajectory based on the prediction result so as to avoid approaching road users. The trajectory generation block 150 outputs the generated trajectory plan to the actuator ECU that controls the driving actuator of the host vehicle A. The host vehicle A is driven in the autonomous driving mode by driving control along the future trajectory by the actuator ECU.

The flow of the behavior prediction method (hereinafter, referred to as the behavior prediction flow) in which the behavior prediction system 100 predicts the behavior of road users through the cooperation of blocks 110, 120, 130, 140, and 150 described above will be described below with reference to FIG. 6. This processing flow is executed repeatedly while the host vehicle A is running. Note that each "S" in this processing flow represents multiple steps executed by multiple commands included in the behavior prediction program.

First, in S10, the road user acquisition block 110 acquires road user identification information that identifies the road user. Next, in S20, the structure acquisition block 120 acquires structure identification information that identifies the structure that the road user can enter and exit.

In the next step S30, the input information preparation block 130 prepares a portion of the input information. Specifically, in S30, structure mapping information, which is map information associated with structure-specific information, is prepared. In addition, in S30, correlation information regarding the correlation between road users and structures as the road users' destinations is acquired.

Then, in S40, the behavior prediction block 140 predicts the behavior of the road user based on the road user identification information, structure mapping information, and correlation information. Specifically, the above information is input into the behavior prediction model, and the behavior prediction result is output. In S50, the future trajectory of the host vehicle A is generated based on the behavior prediction result.

According to the first embodiment described above, the behavior of a road user is predicted based on road user identification information, structure identification information, and correlation information between the road user and the structure. As a result, the correlation between the road user and the structure as the destination is taken into account when predicting the behavior of the road user. Therefore, the behavior of a road user who has a structure as his/her destination can be included in the prediction result. Therefore, the reliability of the behavior prediction can be improved.

Furthermore, according to the first embodiment, the structure identification information includes information regarding the type of structure. Therefore, depending on the type of structure, it may be possible to predict the behavior of the corresponding road user who has the structure as his/her destination. Therefore, it may be possible to make a more reliable prediction of behavior.

Furthermore, according to the first embodiment, the structure identification information includes information regarding the location of the structure. Therefore, depending on the location of the structure, it may be possible to predict the behavior of the corresponding road user who has the structure as his/her destination. Therefore, it may be possible to make a more reliable prediction of the behavior.

In addition, according to the first embodiment, the structure identification information includes information about the entrance and exit of the structure. Therefore, depending on the entrance and exit of the structure, it may be possible to predict the behavior of the corresponding road user with the structure as their destination. Therefore, more reliable behavior prediction may be possible.

Furthermore, according to the first embodiment, the structure identification information includes information on the structure user who uses the structure. Therefore, depending on the characteristics of the structure user, it may be possible to predict the behavior of the corresponding road user who has the structure as his/her destination. Therefore, more reliable behavior prediction may be possible.

Furthermore, predicting the behavior of the road user includes predicting the behavior of the road user based on map information associated with the structure identification information. According to this, the structure identification information is used for behavior prediction in the form of map information. Therefore, the structure identification information can be used in a form suitable as input information for a behavior prediction model such as a trained model. Furthermore, since at least the type of structure is associated with the map information, more reliable behavior prediction according to the type of structure can be made possible.

Second Embodiment
In the second embodiment, a modified example of the behavior prediction system 100 in the first embodiment will be described.

In the second embodiment, the input information preparation block 130 prepares map information associated with a user distribution (structure user distribution) estimated or measured for each structure as structure mapping information. The structure mapping information prepares map information associated with a structure user distribution for each type of road user. In this case, the structure user distribution for each structure corresponds to the type of road user. That is, when the road user is a person, the structure user distribution for each structure according to the attributes is associated with the map information. And, when the road user is a vehicle, the structure user distribution for each structure according to the attributes of the occupant for each type of vehicle is associated with the map information. Therefore, this structure mapping information becomes data that stores correlation information between road users and structures as the destinations of the road users.

In the example shown in FIG. 7, the structure mapping information includes two-dimensional map data associated with a user distribution corresponding to passenger cars, two-dimensional map data associated with a user distribution corresponding to children, two-dimensional map data associated with a user distribution corresponding to ambulances, etc. In FIG. 7, the density of dot hatching increases for structures with a higher distribution of structure users.

The input information preparation block 130 may estimate the structure user distribution based on the characteristic information, etc. Alternatively, the input information preparation block 130 may associate the structure user distribution itself acquired as structure specific information by the structure acquisition block 120 with the map information. Alternatively, the input information preparation block 130 may prepare the map information itself acquired as structure specific information and previously associated with the structure user distribution as structure mapping information.

The behavior prediction block 140 in the second embodiment predicts the behavior of the road user based on the road user identification information and the structure mapping information that stores the correlation. Specifically, the behavior prediction block 140 predicts the behavior by inputting each piece of information into a trained model that has been trained to output a predicted result of the road user's behavior in response to the input of the above information. That is, in the second embodiment, the behavior prediction block 140 does not need to input correlation information separate from the structure mapping information into the behavior prediction model.

In the second embodiment described above, in S30 of the behavior prediction flow in FIG. 6, the input information preparation block 130 prepares map information (structure mapping information) associated with the structure user distribution for each structure as input information. Then, in S40, the behavior prediction block 140 inputs the road user identification information and the structure mapping information including correlation information into the behavior prediction model and obtains the output of the behavior prediction result, thereby executing the behavior prediction. The other steps are processed in the same manner as in the first embodiment.

According to the second embodiment described above, behavior is predicted based on map information to which at least the structure user distribution is associated as structure specific information. Therefore, it is possible to predict behavior based on map information incorporating correlation information. In other words, there is no need to prepare correlation information separate from the map information for behavior prediction. Therefore, the amount of calculation in the behavior prediction process after the map information is generated can be reduced.

Other Embodiments
Although one embodiment has been described above, the present disclosure should not be construed as being limited to the embodiment described above, and can be applied to various embodiments within the scope not departing from the gist of the present disclosure.

In a modified example, the input information preparation block 130 may prepare both map information associated with the type of structure and map information associated with the structure user distribution according to the type of road user as the structure mapping information. In this case, the behavior prediction block 140 may select the structure mapping information to be used for the behavior prediction according to the conditions. For example, the behavior prediction block 140 may be configured to use the map information associated with the structure user distribution according to the type of road user for the behavior prediction when the number of road users to be the target of the behavior prediction is within a set range. In this case, the behavior prediction block 140 may use the map information associated with the type of structure for the behavior prediction when the number of road users to be the target of the behavior prediction is outside the set range. The set range here is a numerical range that is less than or equal to a preset threshold value.

In a variant, the structure mapping information may be three-dimensional map data in which the structure-specific information is expanded.

In a modified example, the behavior prediction system 100 may not generate a future trajectory of the host vehicle A. In other words, the behavior prediction system 100 may not construct a trajectory generation block 150. For example, the behavior prediction system 100 may be configured to output the behavior prediction results to another in-vehicle device or an external server. Alternatively, the behavior prediction system 100 may use the behavior prediction results for processing other than generating a future trajectory, such as presenting information to occupants.

In a modified example, the behavior prediction system 100 may predict the behavior of road users existing in a specific range, rather than predicting the behavior of road users existing in the vicinity of the host vehicle A. In other words, the behavior prediction system 100 may be provided in a system that does not distinguish a specific vehicle as the host vehicle A.

In a modified example, the dedicated computer constituting the behavior prediction system 100 may have at least one of a digital circuit and an analog circuit as a processor. Here, the digital circuit is at least one of the following: ASIC (Application Specific Integrated Circuit), FPGA (Field Programmable Gate Array), SOC (System on a Chip), PGA (Programmable Gate Array), and CPLD (Complex Programmable Logic Device). Such a digital circuit may also have a memory that stores a program.

In addition to the above-described embodiments and variations, the behavior prediction system 100 according to the above-described embodiments and variations may be implemented as a behavior prediction device (e.g., a processing device such as a processing ECU) mounted on the host vehicle A. The above-described embodiments and variations may also be implemented as a semiconductor device (e.g., a semiconductor chip) having at least one processor 102 and one memory 101 of the behavior prediction system 100.

100: Behavior prediction system, 101: Memory (storage medium), 102: Processor

Claims (10)

A behavior prediction system for predicting a behavior of a road user, the behavior prediction system having a processor (102),
The processor,
Obtaining road user identification information for identifying the road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
preparing correlation information between the road users and the structures as destinations of the road users;
predicting a behavior of the road user based on the road user identification information, the structure identification information, and the correlation information;
is configured to run
preparing the correlation information includes preparing map information, for each type of road user, in which a structure user distribution estimated for each of the structures is at least associated as the structure identification information;
A behavior prediction system , wherein predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and the map information for each type of road user . The behavior prediction system of claim 1, wherein the structure identification information includes information regarding the type of the structure. The behavior prediction system according to claim 1 or 2, wherein the structure identification information includes information regarding the position of the structure. The behavior prediction system according to any one of claims 1 to 3, wherein the structure specific information includes information regarding the entrance and exit of the structure. The behavior prediction system according to any one of claims 1 to 4, wherein the structure identification information includes information about a structure user who uses the structure. The behavior prediction system according to claim 1 , wherein the map information is associated with at least a type of the structure. The behavior prediction system according to any one of claims 1 to 6, wherein predicting the behavior of the road user includes inputting the road user identifying information, the structure identifying information, and the correlation information into a trained model trained to output a behavior predicted for the road user in response to input of the road user identifying information, the structure identifying information, and the correlation information. A behavior prediction device for predicting a behavior of a road user, the behavior prediction device having a processor (102),
The processor,
Obtaining road user identification information for identifying the road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
Preparing correlation information between the road users and the structures as destinations of the road users;
predicting a behavior of the road user based on the road user identification information, the structure identification information, and the correlation information;
is configured to run
preparing the correlation information includes preparing map information, for each type of road user, in which a structure user distribution estimated for each of the structures is at least associated as the structure identification information;
A behavior prediction device , wherein predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and the map information for each type of road user . A behavior prediction method executed by a processor (102) for predicting a behavior of a road user, the method comprising:
Obtaining road user identification information for identifying the road user; and obtaining structure identification information for identifying a structure that the road user can enter and exit.
preparing correlation information between the road users and the structures as destinations of the road users;
predicting a behavior of the road user based on the road user identification information, the structure identification information, and the correlation information;
Including,
preparing the correlation information includes preparing map information, for each type of road user, in which a structure user distribution estimated for each of the structures is at least associated as the structure identification information;
The behavior prediction method includes predicting the behavior of the road user based on the road user identification information and the map information for each type of road user . A behavior prediction program stored in a storage medium (101) and including instructions to be executed by a processor (102) for predicting the behavior of road users, the behavior prediction program comprising:
The instruction:
Obtaining road user identification information that identifies the road user;
Obtaining structure identification information that identifies a structure that the road user can enter and exit;
preparing correlation information between the road users and the structures as destinations of the road users;
predicting the behavior of the road user based on the road user identification information, the structure identification information, and the correlation information;
Including,
preparing the correlation information includes preparing map information, in which the structure user distribution estimated for each of the structures is at least associated as the structure identification information, for each type of road user;
A behavior prediction program , wherein predicting the behavior of the road user includes predicting the behavior of the road user based on the road user identification information and the map information for each type of the road user .

Images