CN109697875B

CN109697875B - Method and device for planning driving track

Info

Publication number: CN109697875B
Application number: CN201710993734.8A
Authority: CN
Inventors: 古强; 姚骏
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2017-10-23
Filing date: 2017-10-23
Publication date: 2020-11-06
Anticipated expiration: 2037-10-23
Also published as: CN109697875A

Abstract

The application relates to artificial intelligence, discloses a method and a device for planning a driving track, and belongs to the fields of data processing technology, automatic driving, Internet of things and the like. The method comprises the following steps: the server determines a target neural network model corresponding to a road section where a target position where the first vehicle is located, and sends the target neural network model to the first vehicle, so that the first vehicle determines a driving track through the target neural network model. The target neural network model is obtained by the server through training according to the running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through a road section where the target position where the first vehicle is located currently is located before the current time. That is, when the first vehicle determines its own travel track at the target position through the target neural network model, the travel tracks of the plurality of second vehicles that pass through the road section where the target position is located before the current time are referred to, and the feasibility of the determined travel track is improved.

Description

Method and device for planning driving track

Technical Field

The application relates to the fields of artificial intelligence, data processing technology, automatic driving, Internet of things and the like, in particular to a method and a device for planning a driving track.

Background

The automatic driving, that is, during the vehicle driving, the vehicle plans the driving track and drives according to the planned driving track. However, in practical applications, if the planned driving trajectory of the vehicle is not appropriate, it is easy to affect the safety of other vehicles and the vehicle itself, and therefore how to plan the driving trajectory is very important.

In the related art, when a driving track needs to be planned, a vehicle determines a driving task to be executed and a lane position on a current road, and plans the driving track according to the determined driving task and the lane position on the current road. The driving tasks comprise straight driving, left turning, right turning and turning around, and the lane positions of the current road comprise a straight driving lane, a left turning lane and a right turning lane. For example, the determined driving task is a right turn, the lane position on the current road is a straight lane, and the planned driving trajectory may be: and starting from the center line of the current lane, and driving along the connecting line between the center line of the current lane and the center line of the right lane until the center line of the right lane is reached.

However, when the vehicle travels along the travel track determined by the above method, accidents are easily generated, that is, the accident occurrence rate is high when the vehicle travels along the travel track determined by the above method, and thus the feasibility of the travel track planned by the above method is low.

Disclosure of Invention

In order to solve the problem of low feasibility of a planned driving track in the related art, the application provides a method and a device for planning the driving track. The technical scheme is as follows:

in a first aspect, a method for planning a driving trajectory is provided, which is applied to a first vehicle, and comprises:

receiving a target neural network model sent by a server, wherein the target neural network model is a neural network model corresponding to a road section where a target position where the first vehicle is located, the target neural network model is obtained by the server through training according to the running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through the road section where the target position is located before the current time;

determining a driving track of the first vehicle through the target neural network model according to the driving task and the driving information of the first vehicle and the driving information of the first obstacle vehicle;

the driving task comprises straight driving, left turning, right turning and turning around, the driving information comprises position information of a current position, a driving direction and a driving speed, and the first obstacle vehicle is a vehicle, wherein the distance between the first obstacle vehicle and the first vehicle is smaller than a preset distance threshold value.

In the present application, the target neural network model is obtained by the server through training according to the driving tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through a road section where the target position where the first vehicle is located currently is located before the current time. That is, when the first vehicle determines its own travel track at the target position through the target neural network model, the travel information of the first vehicle and the first obstacle vehicle is considered, and the travel tracks of the plurality of second vehicles passing through the road section where the target position is located before the current time are also referred to, so that the accident rate of the first vehicle traveling according to the determined travel track is reduced, that is, the feasibility of the determined travel track is improved.

Optionally, the road section where the target position is located is an intersection;

the determining, by the target neural network model, a travel trajectory of the first vehicle according to the travel task and the travel information of the first vehicle and the travel information of the first obstacle vehicle includes:

and taking the running task and the running information of the first vehicle, the running information of the first obstacle vehicle and the signal lamp state corresponding to the target position at the intersection as the input of the target neural network model, and determining the running track of the first vehicle through the target neural network model.

Specifically, when the road segment where the target position is located is an intersection, and the driving track of the first vehicle is determined through the target neural network model at this time, the state of a signal lamp corresponding to the target position at the intersection needs to be considered, so that the feasibility of the determined driving track is further improved.

Optionally, the target neural network model is a neural network model corresponding to a road segment where the target position is located and a driving task of the first vehicle, and the plurality of second vehicles are vehicles which pass through the road segment where the target position is located before the current time and have the same driving task as the driving task of the first vehicle.

Further, at a certain road segment, the server may train different neural network models for different driving tasks, and at this time, the first vehicle may determine the driving track according to the target neural network model corresponding to the road segment where the target position is located and the driving task of the first vehicle.

In a second aspect, another method for planning a driving trajectory is provided, and applied to a server, the method includes:

determining a target neural network model corresponding to a road section where a target position where a first vehicle is located from stored neural network models, wherein the target neural network model is obtained by training according to running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through the road section where the target position is located before the current time;

and sending the target neural network model to the first vehicle so that the first vehicle determines the running track of the first vehicle through the target neural network model.

In this application, when the first vehicle travels to the target position, the server may directly determine a target neural network model corresponding to a road segment where the target position is located, and send the target neural network model to the first vehicle, so that the first vehicle determines a travel track through the target neural network model. The target neural network model is obtained by the server through training according to the running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through the road section where the target position is located before the current time. That is, when the first vehicle determines its own travel track at the target position through the target neural network model, the travel tracks of the plurality of second vehicles passing through the road section where the target position is located before the current time are referred to, so as to reduce the accident rate when the first vehicle travels according to the determined travel track, that is, improve the feasibility of the determined travel track.

Optionally, before determining, from the stored neural network models, a target neural network model corresponding to a road segment where a target position where the first vehicle is currently located is located, the method further includes:

determining the driving tracks of all second vehicles passing through the road section where the target position is located in a preset time period and the grade of the driving track of each second vehicle;

selecting N driving tracks with scores larger than a preset score from all the obtained driving tracks, wherein N is larger than 1 and smaller than or equal to the total number of the obtained driving tracks;

and training the initialized neural network model through the N driving tracks to obtain the target neural network model.

Since the first vehicle determines the driving track according to the target neural network model sent by the server, in the present application, the server also needs to determine the target neural network model in advance. Further, in order to determine the feasibility of the driving trajectory through the target neural network model, the server may select an excellent driving trajectory from the plurality of driving trajectories according to the score of each driving trajectory, and train the target neural network model through the selected excellent driving trajectory.

Optionally, the training the initialized neural network model through the N driving trajectories to obtain the target neural network model includes:

determining the running tasks and the running information of the N second vehicles and the running information of the N second obstacle vehicles;

the N second vehicles are vehicles corresponding to the N running tracks, the N second obstacle vehicles correspond to the N second vehicles one by one, the distance between each second obstacle vehicle and the corresponding second vehicle is smaller than a preset distance threshold value, the running tasks comprise straight running, left turning, right turning and turning around, and the running information comprises the current position, the running direction and the running speed;

and training the initialized neural network model according to the running tasks and the running information of the N second vehicles, the running information of the N second obstacle vehicles and the running tracks of the N second vehicles to obtain the target neural network model.

The target neural network model is obtained by determining sample data and then training the initialized neural network model through the determined sample data.

after determining the driving tasks and the driving information of the N second vehicles and the driving information of the N second obstacle vehicles, the method further includes:

determining N signal lamp states, wherein the N signal lamp states correspond to the N second vehicles one by one, and each signal lamp state refers to a corresponding signal lamp state at the intersection when the corresponding second vehicle passes through the intersection;

correspondingly, the training the initialized neural network model according to the driving tasks and the driving information of the N second vehicles, the driving information of the N second obstacle vehicles, and the driving tracks of the N second vehicles to obtain the target neural network model includes:

and taking the running tasks and the running information of the N second vehicles, the running information of the N second obstacle vehicles and the N signal lamp states as the input of the initialized neural network model, taking the running tracks of the N second vehicles as the output of the initialized neural network model, and training the initialized neural network model to obtain the target neural network model.

Further, when the road section where the target position is located is an intersection, the server determines sample data for training the target neural network model, and the sample data may further include a signal lamp state corresponding to the intersection when each second vehicle passes through the intersection.

Optionally, the determining the score of the driving track of each second vehicle includes:

for any one of all second vehicles, determining the running condition of the second vehicle in the running process according to the running track of the second vehicle, wherein the running condition comprises the number of times of collision, whether the traffic regulation is obeyed, the number of lane change, the running time and whether the driving is smooth;

and determining the score of the running track of the second vehicle according to the running condition of the second vehicle in the running process.

The server can determine the score of the driving track of the second vehicle according to the number of times of collision of the second vehicle when the second vehicle passes through the intersection, whether the second vehicle complies with traffic rules, the number of times of lane change, the driving time length, whether the second vehicle is in smooth driving and the like.

Optionally, the determining the running condition of the second vehicle during running according to the running track of the second vehicle includes:

determining a running track of a third obstacle vehicle passing through the intersection, wherein the third obstacle vehicle is a vehicle of which the distance from the second vehicle is smaller than a preset distance threshold when the second vehicle passes through the intersection;

determining the number of collisions between the second vehicle and the third obstacle vehicle according to the travel track of the second vehicle and the travel track of the third obstacle vehicle;

determining a signal lamp state corresponding to the intersection in the process that the second vehicle passes through the intersection;

determining whether the second vehicle complies with traffic rules according to the running track of the second vehicle and the signal lamp state of the intersection in the process that the second vehicle passes through the intersection;

and determining the lane change times, the driving time length and whether the driving is smooth or not of the second vehicle according to the driving track of the second vehicle.

Specifically, the server may determine the number of times of collision of the second vehicle while passing through the intersection, whether to comply with traffic regulations, the number of lane changes, the travel time period, and whether to drive smoothly by the above-described method.

Optionally, the determining a score of the driving track of the second vehicle according to the driving condition of the second vehicle during driving includes:

if the number of times of collision of the second vehicle in the running process is greater than or equal to the preset number of times of collision, determining the collision score as a first score, otherwise, determining the collision score as a second score, wherein the collision score and the number of times of collision are in a negative correlation relationship;

determining a traffic rule score as a third score if the second vehicle complies with traffic rules, otherwise determining the traffic rule score as a fourth score;

if the lane change times of the second vehicle in the driving process are larger than or equal to the minimum lane change times required by the second vehicle to pass through the intersection, determining that the lane change score is a fifth score, and if not, determining that the lane change score is a sixth score, wherein the lane change score and the lane change times are in a negative correlation relationship;

if the running time of the second vehicle passing through the intersection is greater than or equal to the preset running time, determining that the time score is a seventh score, otherwise, determining that the time score is an eighth score, wherein the time score and the running time are in a negative correlation relationship;

determining a driving score as a ninth score if the second vehicle is driving smoothly, otherwise determining the driving score as a tenth score;

determining the sum of the collision score, the traffic regulation score, the lane change score, the duration score and the driving score as the score of the driving track of the second vehicle.

Further, when determining the number of times of collision, whether the second vehicle obeys traffic rules, the number of times of lane change, the driving time and whether the second vehicle is in stable driving, the server may determine a collision score, a traffic rule score, a lane change score, a time score and a driving score, respectively, so as to determine a score of the driving trajectory of the second vehicle according to the determined collision score, traffic rule score, lane change score, time score and driving score.

Optionally, the determining, from the stored neural network models, a target neural network model corresponding to a road segment where a target position where the first vehicle is currently located includes:

determining a target neural network model corresponding to a road section where the target position is located and the driving task of the first vehicle from stored neural network models according to the position information of the target position and the driving task of the first vehicle;

correspondingly, the plurality of second vehicles are vehicles which pass through the road section where the target position is located before the current time and have the same running task as that of the first vehicle.

Further, at a certain road segment, the server may train different neural network models for different driving tasks, and at this time, the server may send, to the first vehicle, a target neural network model corresponding to both the road segment where the target position is located and the driving task of the first vehicle.

Optionally, the determining the driving tracks of all second vehicles passing through the road segment where the target position is located within the preset time period includes:

determining a second vehicle passing through a road section where the target position is located at a first moment, wherein the first moment is any moment in the preset time period;

for any one determined second vehicle, determining a plurality of second moments when the second vehicle passes through the road section where the target position is located in the driving process;

determining a travel track of the second vehicle based on the travel information of the second vehicle at each second time.

The server determines the driving track of the second vehicle, that is, determines the driving information of the second vehicle at each second moment in the process of passing through the road section where the target position is located.

In a third aspect, a device for planning a driving track is provided, which is applied to a first vehicle, and has the function of implementing the method for planning a driving track in the first aspect. The device for planning a driving trajectory comprises at least one unit for implementing the method for planning a driving trajectory provided in the first aspect.

In a fourth aspect, another apparatus for planning a driving trajectory is provided, which is applied to a server, and has a function of implementing the method for planning a driving trajectory in the second aspect. The device for planning a driving trajectory comprises at least one unit for implementing the method for planning a driving trajectory according to the second aspect.

In a fifth aspect, an apparatus for planning a driving trajectory is provided, where the apparatus for planning a driving trajectory structurally includes a processor and a memory, and the memory is used for storing a program supporting a beamforming apparatus to execute the method for planning a driving trajectory provided in the first aspect, and storing data related to implementing the method for planning a driving trajectory provided in the first aspect. The processor is configured to execute programs stored in the memory. The operating means of the memory device may further comprise a communication bus for establishing a connection between the processor and the memory.

In a sixth aspect, another apparatus for planning a driving trajectory is provided, where the apparatus for planning a driving trajectory includes a processor and a memory, and the memory is used to store a program for supporting a beamforming apparatus to execute the method for planning a driving trajectory provided in the second aspect, and store data for implementing the method for planning a driving trajectory provided in the second aspect. The processor is configured to execute programs stored in the memory. The operating means of the memory device may further comprise a communication bus for establishing a connection between the processor and the memory.

In a seventh aspect, a computer-readable storage medium is provided, which has instructions stored therein, which when run on a computer, cause the computer to perform the method for planning a driving trajectory according to the first aspect.

In an eighth aspect, another computer-readable storage medium is provided, having stored therein instructions, which when run on a computer, cause the computer to execute the method of planning a driving trajectory according to the second aspect described above.

In a ninth aspect, there is provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the method of planning a driving trajectory of the first aspect described above.

In a tenth aspect, there is provided another computer program product comprising instructions which, when run on a computer, cause the computer to perform the method of planning a driving trajectory according to the second aspect described above.

The technical effects obtained by the third, fifth, seventh and ninth aspects are similar to the technical effects obtained by the corresponding technical means in the first aspect, and are not described herein again. The technical effects obtained by the above fourth, sixth, eighth and tenth aspects are similar to the technical effects obtained by the corresponding technical means in the second aspect, and are not repeated here again.

The beneficial effect that technical scheme that this application provided brought is:

in this application, when the first vehicle is currently located at the target position, the server may directly determine a target neural network model corresponding to a road segment where the target position is located, and send the target neural network model to the first vehicle, so that the first vehicle determines the driving track through the target neural network model. The target neural network model is obtained by the server through training according to the running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through a road section where the target position where the first vehicle is located currently is located before the current time. That is, when the first vehicle determines its own travel track at the target position through the target neural network model, the travel information of the first vehicle and the first obstacle vehicle is considered, and the travel tracks of the plurality of second vehicles passing through the road section where the target position is located before the current time are also referred to, so that the accident rate of the first vehicle traveling according to the determined travel track is reduced, that is, the feasibility of the determined travel track is improved.

Drawings

FIG. 1 is a schematic diagram of a system for planning a driving trajectory according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a server according to an embodiment of the present invention;

fig. 3 is a schematic structural diagram of a server according to an embodiment of the present invention;

FIG. 4 is a schematic illustration of a vehicle according to an embodiment of the present invention;

FIG. 5 is a schematic structural diagram of a vehicle according to an embodiment of the present invention;

fig. 6A is a flowchart of a method for planning a driving trajectory according to an embodiment of the present invention;

FIG. 6B is a schematic view of an intersection according to an embodiment of the present invention;

FIG. 7 is a flowchart of another method for planning a driving trajectory according to an embodiment of the present invention;

FIG. 8 is a block diagram of an apparatus for planning a driving trajectory according to an embodiment of the present invention;

FIG. 9A is a block diagram of an apparatus for planning a driving trajectory according to an embodiment of the present invention;

fig. 9B is a block diagram of another apparatus for planning a driving trajectory according to an embodiment of the present invention.

Detailed Description

To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.

For ease of understanding, first, a brief description is made of an application scenario related to the embodiment of the present invention.

For an automatically-driven vehicle, the vehicle runs according to a pre-planned running track, but not according to an instruction issued by a driver in real time like a driver-driven vehicle, so that the automatically-driven vehicle cannot adaptively adjust the running direction or the running speed according to the current environment of the vehicle in real time like the driver-driven vehicle. For example, the current driving task of the vehicle is right turn, the lane position on the current road is a straight lane, and the planned driving track at this time is: and starting from the center line of the current lane, and driving along the connecting line between the center line of the current lane and the center line of the right lane until the center line of the right lane is reached. In the process of right-turn driving of the vehicle, other vehicles may have lane-changing behaviors around the vehicle, and at this time, if the vehicle continues to drive according to the planned driving track, the other vehicles and the safety of the vehicle are easily affected. Therefore, it is important to plan the driving trajectory. The embodiment of the invention is applied to a scene of how to plan the driving track of the vehicle.

After the application scenario of the embodiment of the present invention is introduced, a system for planning a driving trajectory provided by the embodiment of the present invention is briefly introduced below.

As shown in fig. 1, an embodiment of the present invention provides a system for planning a driving trajectory, where the system 100 for planning a driving trajectory includes a server 101 and at least one vehicle 102, and the server 101 and each vehicle 102 are wirelessly connected for communication.

The server 101 is configured to count the historical driving tracks, so as to train the neural network model through the historical driving tracks. The vehicle 102 is configured to obtain the pre-trained neural network model from the server 101 and plan a driving track through the pre-trained neural network.

Optionally, the server 101 is further configured to push the pre-trained neural network model to at least one vehicle 102 in a broadcast manner, where the vehicle 102 does not need to actively obtain the pre-trained neural network model from the server 101.

Fig. 2 is a schematic diagram of a server 200 according to an embodiment of the present invention. Referring to fig. 2, the server 200 includes a data collection module 201, a data storage module 202, a task analysis module 203, a time trajectory analysis module 204, a signal light analysis module 205, a trajectory scoring module 206, a model training module 207, and a model push module 208.

The data acquisition module 201 is configured to acquire data, where the acquired data may be data reported by a plurality of fixedly installed cameras installed on a preset road segment, or data reported to a server by a vehicle passing through the preset road segment. If the acquired data is data reported by a plurality of cameras, the acquired data at least comprises: the video that gathers, the time that this video corresponds, the geographical position of camera, the information such as the height of camera and the orientation of camera. If the collected data is data reported by vehicles passing through the preset road section, the collected data at least comprises the following data: the vehicle is located at a position, a driving direction, a speed and corresponding time information. Optionally, the collected data may further include a video shot by a vehicle-mounted camera installed on the vehicle, or information after processing the shot video through vehicle-mounted software installed on the vehicle. The video shot by the vehicle-mounted camera can comprise signal light information, and information such as the position, the driving direction and the speed of a peripheral dynamic obstacle.

The data storage module 202 is used for storing the data acquired by the data acquisition module 201.

The task analysis module 203 is configured to read the stored data from the data storage module 202, determine a driving track of each vehicle through a preset video processing technology, and analyze a driving task of each vehicle on the preset road segment according to the driving track of each vehicle. Such as turning left, going straight, turning right, or turning around, etc.

The time trajectory analysis module 204 reads the stored data from the data storage module 202, and determines a driving trajectory of each vehicle through a preset video processing technology, where the driving trajectory of the vehicle includes information such as a position, a driving direction, and a vehicle speed of the vehicle in a time sequence, that is, the driving trajectory of the vehicle is composed of information such as the position, the driving direction, and the vehicle speed of the vehicle when the vehicle drives on the preset road segment at different times.

The traffic light analysis module 205 reads the stored data from the data storage module 202, and similarly analyzes information of the traffic lights of the preset section with time through a preset video processing technique. For example, it is analyzed that the straight traffic light in the west-east direction is green in the time period t0 to t 1.

The trajectory scoring module 206 is configured to read data processed by the task analysis module 203, the time trajectory analysis module 204, and the signal light analysis module 205, correlate the data with each other according to time, and regard different vehicles within a preset distance threshold as vehicles that are obstacles to each other, so as to score all driving trajectories, where the scoring criteria may include: whether to comply with traffic regulations, time to pass through the preset section, number of lane changes, number of collisions with obstacle vehicles, etc.

The model training module 207 trains the historical driving track by using the initialized neural network model according to the scoring data of the task analysis module 203, the time track analysis module 204, the signal lamp analysis module 205 and the track scoring module, so that the neural network model learns the characteristics of the driving track with higher score to obtain the neural network model, and stores the trained neural network model.

The model pushing module 208 is configured to push the stored neural network model to the vehicle when a model request of the vehicle is received, and the model pushing module 208 is further configured to push the neural network model to a range of vehicles in a broadcast manner.

Fig. 3 is a schematic structural diagram of another server 300 according to an embodiment of the present invention. The server 300 is used to implement the functions of the respective modules included in the server 200 shown in fig. 2. Specifically, as shown in FIG. 3, the server 300 includes at least one processor 301, a communication bus 302, a memory 303, and at least one communication interface 304.

The processor 301 may be a general-purpose Central Processing Unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more ics for controlling the execution of programs in accordance with the present invention.

The communication bus 302 may include a path that conveys information between the aforementioned components.

The Memory 303 may be, but is not limited to, a Read-Only Memory (ROM) or other type of static storage device that can store static information and instructions, a Random Access Memory (RAM) or other type of dynamic storage device that can store information and instructions, an Electrically Erasable Programmable Read-Only Memory (EEPROM), a compact disc Read-Only Memory (CD-ROM) or other optical disc storage, optical disc storage (including compact disc, laser disc, optical disc, digital versatile disc, blu-ray disc, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. The memory 303 may be separate and coupled to the processor 301 through a communication bus 302. The memory 303 may also be integrated with the processor 301.

Communication interface 304, using any transceiver or the like, is used for communicating with other devices or communication Networks, such as ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

In particular implementations, processor 301 may include one or more CPUs, as one embodiment.

In particular implementations, a server may include multiple processors, as one embodiment. Each of these processors may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions). For example, for each of the task analysis module 203, the time trajectory analysis module 204, the signal light analysis module 205, the trajectory scoring module 206, and the model training module 207 shown in fig. 2, a processor may implement the functions of the module.

In one implementation, the server 300 may further include an output device and an input device. An output device, which is in communication with the processor 301, may display information in a variety of ways. For example, the output device may be a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display device, a Cathode Ray Tube (CRT) display device, a projector (projector), or the like. The input device is in communication with the processor 301 and may receive user input in a variety of ways. For example, the input device may be a mouse, a keyboard, a touch screen device, or a sensing device, among others.

In a specific implementation, the server may be a desktop, a laptop, a network server, a Personal Digital Assistant (PDA), a mobile phone, a tablet, a wireless terminal device, a communication device, or an embedded device. The embodiment of the invention does not limit the type of the server.

The memory 303 is used for storing program codes for executing the scheme of the application, and is controlled by the processor 301 to execute so as to train the neural network model according to the historical driving track. The processor 301 is operable to execute program code stored in the memory 303. One or more software modules may be included in the program code.

Fig. 4 is a schematic diagram of a vehicle 400 according to an embodiment of the present invention, where the vehicle 400 includes a positioning module 401, a model request module 402, an in-vehicle communication module 403, a mission planning module 404, a sensing module 405, a trajectory calculation module 406, and a vehicle control module 407.

The positioning module 401 collects the current position information of the vehicle in real time, and sends the collected position information to the model requesting module 402 at a certain frequency.

The model requesting module 402 is configured to request the neural network model from the server through the vehicle-mounted communication module 403, that is, the model requesting module 402 is configured to send a model obtaining request to the server through the vehicle-mounted communication module 403, where the model obtaining request includes a target position where the vehicle is currently located. When the model request module 402 receives the neural network model returned by the server through the in-vehicle communication module 403, the neural network model is stored.

Alternatively, when the model pushing module 208 in the server shown in fig. 2 pushes the neural network model to a certain range of vehicles by broadcasting, the model requesting module 402 is configured to receive the neural network model pushed by the server directly through the vehicle-mounted communication module 403, and store the neural network model.

In addition, when different road sections correspond to different neural network models, the model request module 402 is further configured to check whether a local neural network model corresponding to the current driving road section exists according to the acquired location information, and when the neural network model does not exist locally, obtain the neural network model through the server.

The mission planning module 404 is used to determine a driving mission for the vehicle. The sensing module 405 is used to determine information such as the traveling information of the vehicle, the traveling information of the obstacle vehicle, and the state of the traffic light. The driving information includes information such as a current position, a driving direction, and a driving speed.

The trajectory calculation module 406 is configured to determine a driving trajectory of the vehicle through the neural network model obtained by the model request module 402 according to the data determined by the mission planning module 404 and the sensing module 405. The vehicle control module 407 is configured to control the vehicle to travel according to the travel track determined by the track calculation module 406.

Fig. 5 is a schematic structural diagram of another vehicle 500 according to an embodiment of the present invention, where the vehicle 500 is used to implement the functions of the modules included in the vehicle 400 shown in fig. 4. Specifically, as shown in fig. 5, the vehicle 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

The processor 501 and the processor 301 shown in fig. 3 have substantially the same structure and function, the communication bus 502 and the communication bus 502 shown in fig. 3 have substantially the same structure and function, the memory 503 and the memory 503 shown in fig. 3 have substantially the same structure and function, and the communication interface 504 shown in fig. 3 have substantially the same structure and function, and therefore, detailed description thereof is omitted. ,

except that the memory 503 is used for storing program codes for executing the scheme of the application and is controlled by the processor 501 to execute so as to realize the planning of the driving track for the vehicle according to the trained neural network model.

Embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

Fig. 6A is a flowchart of a method for planning a driving track according to an embodiment of the present invention, where the method is applied to the system shown in fig. 1, and for convenience of description, a vehicle that needs to plan a driving track at present is referred to as a first vehicle. As shown in fig. 6A, the method includes the steps of:

step 601: the server determines a target neural network model corresponding to a road section where a target position where the first vehicle is located from the stored neural network models, wherein the target neural network model is obtained by the server through training according to the running tracks of a plurality of second vehicles, and the plurality of second vehicles are vehicles passing through the road section where the target position is located before the current time.

In practical applications, the geographical situation of different road segments may vary greatly, for example, the road smoothness of different road segments and fixed obstacles on the road may be different, so in the embodiment of the present invention, the server may train different neural network models for different road segments in advance. Since different road segments correspond to different neural network models, when the first vehicle is currently at the target position, the server may determine the target neural network model corresponding to the road segment at which the target position is located.

Particularly, because the geographical situation at the intersection is complex, the accident rate of the first vehicle passing through the intersection is high, and the geographical situations of the road sections at other non-intersections are relatively simple, in the embodiment of the invention, the server can train different neural network models only for different intersections. That is, when the road segment where the target position is located is the intersection, the target neural network model is the neural network model corresponding to the intersection.

It should be noted that, in the embodiment of the present invention, the intersection may be all branch intersections at the intersection of two different roads, or may be any branch intersection of all branch intersections at the intersection of the two different roads. For example, as shown in fig. 6B, the road 1 and the road 2 form an intersection at the intersection, and the intersection includes four different branch intersections in four directions, i.e., south, east, west, and north.

When the intersection in the embodiment of the present invention is the intersection, the server determines the neural network model corresponding to the intersection according to the historical driving tracks of all the branch intersections included in the intersection. When the intersection in the embodiment of the present invention can be a branch intersection in any direction in the intersection, the server determines the neural network model corresponding to the branch intersection according to the historical driving track passing through the ten-branch intersection, that is, at the intersection, 4 different neural network models can be trained for each upward branch intersection.

In addition, for the same road section, the difference between the corresponding driving tracks of different driving tasks at the road section is large. In the process of training the neural network, a training sample with a certain unified rule needs to be analyzed, so that in the embodiment of the invention, in order to facilitate the neural network model to learn the rule of data in the training sample, different neural network models can be respectively trained according to the driving task for the same road section. That is, at the road segment, a neural network model is trained for each driving task.

At the moment, when the first vehicle needs to determine the driving track, the server determines that the target neural network model corresponds to the road section where the first vehicle driving task and the target position are located. Correspondingly, when the server trains the target neural network model according to the running tracks of a plurality of second vehicles, the plurality of second vehicles are vehicles which pass through the road section where the target position is located before the current time and have the same running task as that of the first vehicle.

The implementation manner of the server training the target neural network model according to the driving tracks of the plurality of second vehicles will be described in detail in the following embodiments, which will not be described herein.

Step 602: the server sends the target neural network model to the first vehicle.

After determining the target neural network model through step 601, the server transmits the target neural network model to the first vehicle, so that the first vehicle determines the driving track of the first vehicle according to the target neural network model through the following

steps

603 and 604.

Step 603: the first vehicle receives the target neural network model sent by the server.

When the first vehicle receives the target neural network model transmitted by the server, the travel track of the first vehicle can be determined by the following step 604.

Step 604: the first vehicle determines a running track of the first vehicle through the target neural network model according to a running task and running information of the first vehicle and the running information of a first obstacle vehicle, wherein the running task comprises straight running, left turning, right turning and turning around, the running information comprises a current position, a running direction and a running speed, and the first obstacle vehicle is a vehicle of which the distance from the first obstacle vehicle is smaller than a preset distance threshold value.

In the embodiment of the present invention, the server may train the initialized neural network model according to the driving trajectories of all second vehicles passing through the road section where the target position is located within the preset time period, to obtain the target neural network model corresponding to the road section where the target position is located, so that the first vehicle determines the driving trajectory according to the target neural network model. That is, when the first vehicle determines its own travel track at the target position through the target neural network model, the travel information of the first vehicle and the first obstacle vehicle is considered, and the travel tracks of the plurality of second vehicles passing through the road section where the target position is located before the current time are also referred to, so that the accident rate of the first vehicle traveling according to the determined travel track is reduced, that is, the feasibility of the determined travel track is improved.

Fig. 7 is a flowchart of another method for planning a driving trajectory according to an embodiment of the present invention, where the method is applied to the system shown in fig. 1. The embodiment shown in fig. 7 is a further detailed description of the embodiment shown in fig. 6A, and as shown in fig. 7, the method includes the following steps:

step 701: and the server determines the running tracks of all second vehicles passing through the road section where the target position is located in a preset time period, wherein the target position is the current position of the first vehicle.

In the embodiment of the present invention, since the target neural network model is obtained by the server according to the training of the driving trajectories of the plurality of second vehicles, before the server sends the target neural network model to the first vehicle, the server needs to train the driving trajectories of the plurality of second vehicles to determine the target neural network model. Specifically, the target neural network model may be determined through steps 701 to 705.

In addition, in order to facilitate the neural network model to learn the rule of data in the training sample, different neural network models can be trained for the same road section according to the driving task. Since the difference of training different neural network models is that the training samples are different, in the embodiment of the present invention, the example of training the neural network model corresponding to the road segment where the target position is located is taken as an example for description, and the training process of other types of neural network models is not described in detail.

It should be noted that the key point of training the neural network is to determine a training sample, and when the neural network model corresponding to the road segment where the target position is located needs to be trained, the training sample needs to be determined according to the travel track passing through the road segment where the target position is located before the current time.

In addition, because the number of the driving tracks passing through the road section where the target position is located before the current time is possibly large, and the geographical situation of the road section where the target position is located before the current time is also possibly changed, that is, the data of the driving tracks occurring when the distance from the current time is long may be invalid, the neural network model only needs to be trained according to the historical driving tracks in the preset time period.

For example, the preset time period may be 1 month before the current time, and at this time, the neural network model may be determined according to the driving track of the second vehicle passing through the road segment where the target position is located within 1 month before the current time.

Specifically, a second vehicle passing through a road section where the target position is located at a first time is determined, the first time is any time within the preset time period, for any determined second vehicle, a plurality of second times in the running process of the second vehicle passing through the road section where the target position is located are determined, and the running track of the second vehicle is determined based on the running information of the second vehicle at each second time.

In one possible implementation manner, in order to determine all the travel tracks passing through the road section where the target position is located within the preset time period, a plurality of first moments may be divided within the preset time period, and a second vehicle passing through the road section where the target position is located at each first moment is determined, where the second vehicles passing through the road section where the target position is located at all the first moments are all the vehicles passing through the road section where the target position is located within the preset time period.

The running track of each second vehicle is composed of running information of each second time when the second vehicle passes through the intersection, and the running information comprises information such as the current position, running direction and running speed of the second vehicle at the second time.

Specifically, the data storage module shown in fig. 2 determines the video acquired by the data acquisition module at the road section where the target position is located within the preset time period, and determines the driving track of each second vehicle through the time track analysis shown in fig. 2.

In the embodiment of the invention, in order to obtain a feasible driving track through the neural network model, a relatively excellent driving track in historical driving tracks is adopted to train the neural network model. That is, after determining the travel tracks of all the second vehicles that have traveled the section of the target position within the preset time period, the excellent travel track is determined.

When the road segment where the target position is located is an intersection, the server may specifically determine the score of each driving track through the following

steps

702 and 703, so as to determine the excellent driving track according to the score of each driving track through the following step 704.

Alternatively, when the road segment where the target position is located is another type of road segment, the excellent driving track may also be determined by referring to the following steps 702 to 704, which will not be described in detail herein.

Step 702: the server determines the running condition of any one of all the second vehicles in the running process of the second vehicle according to the running track of the second vehicle, wherein the running condition comprises the number of times of collision, whether the traffic regulation is obeyed, the number of lane change, the running time and whether the driving is smooth.

Wherein, the determining of the number of times of collision may specifically be: and determining the running track of a third obstacle vehicle when the third obstacle vehicle passes through the intersection, wherein the distance between the third obstacle vehicle and the second vehicle is less than a preset distance threshold value. Determining the number of collisions between the second vehicle and the third obstacle vehicle based on the travel track of the second vehicle and the travel track of the third obstacle vehicle.

Since the travel track of the second vehicle is composed of the travel information of the second vehicle at each second time when the second vehicle passes through the intersection, the travel track of the third obstacle vehicle is also composed of the travel information of the third obstacle vehicle at each second time, that is, the travel tracks of the second vehicle and the third obstacle vehicle correspond to each other one by one at the time. At this time, the determination of the number of collisions between the second vehicle and the third obstacle vehicle may be achieved by: and for each second moment, determining the current position of the second vehicle and the current position of the third obstacle vehicle, and if the distance between the current position of the second vehicle and the current position of the third obstacle vehicle is less than the preset collision distance, determining that the second vehicle collides with the third obstacle vehicle at the second moment.

The preset collision distance is a preset distance, and the preset collision distance may be 0.1m, 0.15m, 0.2m, or the like.

Secondly, determining whether the second vehicle complies with the traffic rules may specifically be: and determining a signal lamp state corresponding to the intersection in the process that the second vehicle passes through the intersection, and determining whether the second vehicle complies with traffic rules or not according to the running track of the second vehicle and the signal lamp state corresponding to the intersection in the process that the second vehicle passes through the intersection.

That is, in the process that the second vehicle travels at each second time, the signal lamp state corresponding to the intersection at each second time is determined, and for each second time, at the second time, if the signal lamp state corresponding to the intersection is a red lamp, it indicates that the second vehicle does not comply with the traffic rules at that time, and if the signal lamp state corresponding to the intersection is a green lamp, it indicates that the second vehicle complies with the traffic rules at that time.

In addition, according to the running track of the second vehicle, the lane change times, the running time length and whether the second vehicle is in smooth driving are determined.

In order to accurately describe the time length of the second vehicle passing through the intersection, the determining the time length may specifically be: determining the time of the second vehicle passing through the intersection, determining the traffic prohibition time length when the signal lamp is red according to the corresponding signal lamp state in the time, and subtracting the traffic prohibition time length from the time of the second vehicle passing through the intersection to obtain the time which is the running time length.

The determination of whether the second vehicle is driving smoothly may specifically be: and determining the running speed of the second vehicle at each second moment, and performing variance calculation on the running speed of the second vehicle at each second moment to obtain a speed variance. Since the variance may describe the degree of dispersion of a set of data, if the speed variance is greater than the preset variance, it indicates that the speed of the second vehicle is unstable, i.e., the second vehicle is not driving smoothly, and if the speed variance is less than the preset variance, it indicates that the speed of the second vehicle is stable, i.e., the second vehicle is driving smoothly.

Step 703: and the server determines the score of the running track of the second vehicle according to the running condition of the second vehicle in the running process.

In order to determine the superior travel track of all the travel tracks of the second vehicles, the travel track of each second vehicle may be scored. Wherein the scoring criterion is the running condition of the second vehicle in the running process.

Specifically, if the number of times of collision of the second vehicle during the driving process is greater than or equal to a preset number of times of collision, the collision score is determined to be a first score, otherwise, the collision score is determined to be a second score, wherein the collision score and the number of times of collision are in a negative correlation relationship.

If the second vehicle complies with the traffic rules, the traffic rules score is determined to be a third score, otherwise, the traffic rules score is determined to be a fourth score.

And if the lane change times of the second vehicle in the driving process are larger than or equal to the minimum lane change times required by the second vehicle to pass through the intersection, determining that the lane change score is a fifth score, otherwise, determining that the lane change score is a sixth score, wherein the lane change score and the lane change times are in a negative correlation relationship.

And if the running time of the second vehicle passing through the intersection is greater than or equal to the preset running time, determining that the time score is a seventh score, otherwise, determining that the time score is an eighth score, wherein the time score and the running time are in a negative correlation relationship.

If the second vehicle is driving smoothly, the driving score is determined to be a ninth score, otherwise, the driving score is determined to be a tenth score.

And determining the sum of the collision score, the traffic rule score, the lane change score, the duration score and the driving score as the score of the driving track of the second vehicle.

The first score, the second score, the third score, the fourth score, the fifth score, the sixth score, the seventh score, the eighth score, the ninth score and the tenth score are preset scores, and the preset scores can be any scores and only need to meet the conditions.

For example, the second score, the third score, the sixth score, the eighth score, and the ninth score are set to +5 points, and the first score, the fourth score, the fifth score, and the tenth score are set to-5 points. And if the number of times of collision of the second vehicle in the running process is greater than or equal to the preset number of times of collision, determining that the collision score is +5, namely, the score of the running track of the second vehicle is added with 5 at the moment. If the second vehicle does not comply with the traffic regulation, it is determined that the traffic regulation score is-5, i.e., the score of the traveling track of the second vehicle is reduced by 5 at this time.

That is, in the embodiment of the present invention, the driving tracks of all the second vehicles passing through the road section where the target position is located within the preset time period and the score of each driving track may be determined through the above steps 701 to 703.

For example, a plurality of first time instants t0, t1, t2, … and tm are divided within a preset time period, the driving tracks of the second vehicle and the second vehicle passing through the intersection at the first time instant t0 are determined from the first time instant t0, and the score of the driving track is determined through the steps 802 and 803. Continuing to the next time t1, repeating the above process, determining the driving track score of the second vehicle passing through the intersection at the first time t1, …, and so on until determining the driving track score of the second vehicle passing through the intersection at the last first time tm.

Step 704: the server selects N driving tracks with scores larger than a preset score from all the acquired driving tracks, wherein N is larger than 1 and smaller than or equal to the total number of the acquired driving tracks.

Since the score of the travel track is determined according to the travel condition of the second vehicle during the travel, the score of the travel track can be used to describe the excellence degree of the travel track, that is, the higher the score of the travel track, the more excellent the travel track is. Therefore, a relatively excellent travel locus among the historical travel loci can be obtained through step 704.

The preset score is a preset score, and the preset score can be 80, 90 or 95.

Step 705: and the server trains the initialized neural network model through the N driving tracks to obtain a target neural network model.

For the training sample of the neural network model in the embodiment of the present invention, the training sample includes a plurality of independent variables and a plurality of dependent variables corresponding to the plurality of independent variables one to one, for convenience of description, the plurality of independent variables are labeled as x1, x2, …, xn, and the plurality of dependent variables corresponding to the plurality of independent variables one to one are labeled as y1, y2, …, yn. Training the neural network model, that is, enabling the initialized neural network model to learn the mapping relationship between the independent variables and the dependent variables corresponding to the independent variables one by one, and obtaining y ═ f (x), which is the trained neural network model.

Therefore, the step 705 may specifically be determining the running tasks and the running information of the N second vehicles and the running information of the N third obstacle vehicles, where the N second vehicles are vehicles corresponding to the N running tracks, the N second obstacle vehicles are in one-to-one correspondence with the N second vehicles, and the second obstacle vehicles are vehicles whose distance to the corresponding second vehicles is smaller than the preset distance threshold.

And taking the running tasks and the running information of the N second vehicles and the running information of the N second obstacle vehicles as the input of the initialized neural network model, taking the running tracks of the N second vehicles as the output of the initialized neural network model, and training the initialized neural network model to obtain the neural network model corresponding to the road section where the target position is located.

Further, when the road segment where the target position is located is an intersection, the step 705 may specifically be: determining the running tasks and the running information of the N second vehicles, the states of N signal lamps corresponding to the N second vehicles one by one at the intersection and the running information of N second obstacle vehicles

And taking the running tasks and the running information of the N second vehicles, the N signal lamp states and the running information of the N second obstacle vehicles as the input of the initialized neural network model, taking the running tracks of the N second vehicles as the output of the initialized neural network model, and training the initialized neural network model to obtain the neural network model corresponding to the intersection.

That is, the driving tasks and the driving information of the N third vehicles, the signal lamp states of the intersection corresponding to the N second vehicles one to one, and the driving information of the N second obstacle vehicles are used as independent variables in the training sample, and the driving tracks of the N second vehicles are used as dependent variables in the training sample, so as to determine the mapping relationship y ═ f (x) between the independent variables and the dependent variables, that is, determine the neural network model.

Since the data of the training samples is determined from the traveling tracks of all the second vehicles passing through the intersection within the preset time period, the obtained neural network model is the neural network model corresponding to the intersection.

It should be noted that, after the server determines that the server is the neural network model corresponding to the road segment where the target position is located, the server may store the neural network model, that is, store a correspondence between the neural network model and the road segment where the target position is located in the server, that is, one road segment corresponds to one neural network model.

After that, the server may send, to the first vehicle, a target neural network model corresponding to a target location where the first vehicle is currently located according to the model acquisition request sent by the first vehicle. Or the server pushes the neural network model corresponding to the road section to the vehicles passing through the road section in a broadcasting mode.

As can be seen from the above steps 701 to 705, in order to obtain the neural network model corresponding to the road segment where the target position is located, the training sample may be determined according to the driving track passing through the road segment where the target position is located within the preset time period. Therefore, for a certain road segment, if a neural network model corresponding to a certain driving task is desired to be obtained, a training sample can be determined according to a driving track passing through the road segment and the driving task being the driving task, and after the training sample is obtained, a process of training the corresponding neural network model through the training sample is basically the same as the process of training the neural network model, and the embodiment of the present invention is not described in detail herein.

That is, through the above steps 701 to 705, the server may train different neural network models in advance for different road segments, so that the first vehicle determines the driving track through the target neural network model corresponding to the road segment where the current target position is located. Specifically, the travel locus of the first vehicle can be determined by the following steps 706 to 707.

Step 706: the server determines a target neural network model corresponding to a road section where a target position where the first vehicle is located from the stored neural network models.

In a possible implementation manner, when the server receives a model acquisition request sent by a first vehicle, according to a current target location of the first vehicle carried in the model acquisition request, a target neural network model corresponding to a road segment where the target location is located is determined from stored neural network models corresponding to different road segments, and the target neural network model is sent to the first vehicle through the following step 707.

In another possible implementation manner, for the road segment where the target location is located, the server determines, from the stored neural network models corresponding to different road segments, a target neural network model corresponding to the road segment where the target location is located, and pushes the target neural network model to the vehicle in the range where the road segment is located in a broadcast manner through the following step 707. Accordingly, the first vehicle currently at the target location may directly receive the target neural network model.

Step 707: the server sends the target neural network model to the first vehicle so that the first vehicle determines the running track of the first vehicle through the target neural network model.

The first vehicle can determine the driving trajectory from the target neural network model in

steps

708 and 709, which are described below.

Step 708: the first vehicle receives the target neural network model sent by the server.

As can be seen from step 707, in the embodiment of the present invention, the first vehicle may receive the target neural network model from the server through two different implementations. And will not be described in detail herein.

Step 709: the first vehicle determines a running track of the first vehicle through the target neural network model according to a running task and running information of the first vehicle and the running information of a first obstacle vehicle, wherein the running task comprises straight running, left turning, right turning and turning around, the running information comprises a current position, a running direction and a running speed, and the first obstacle vehicle is a vehicle of which the distance from the first obstacle vehicle is smaller than a preset distance threshold value.

Specifically, step 709 can be implemented by the following two steps:

(1) the travel task of the first vehicle, the travel information of the first vehicle, and the travel information of the first obstacle vehicle are acquired.

Specifically, the first vehicle may determine its current driving task through the task planning module shown in fig. 4, so as to obtain the driving task of the first vehicle. The current position of the first obstacle vehicle is determined through the positioning module, the driving direction and the driving speed of the first obstacle vehicle are determined through the sensing module, and the current position, the driving direction, the driving speed and other information of the first obstacle vehicle are obtained.

The task planning module determines whether the current running task is straight, left-turning, right-turning or turning around according to a predetermined navigation path and the current position. The Positioning module may determine a current location of the first vehicle through a Global Positioning System (GPS) technology. The sensing module may determine a driving direction, a driving speed, and driving information of the first obstacle vehicle through a video captured by a camera mounted to the first vehicle.

In addition, the preset distance threshold is a preset distance, and when the distance between two vehicles is smaller than the preset distance threshold, the driving condition of one vehicle may affect the driving condition of the other vehicle, and at this time, the two vehicles are obstacle vehicles. The preset distance threshold may be 1 meter, 0.75 meter, 0.5 meter, or the like.

Optionally, when the road segment where the target position of the first vehicle is located is the intersection, in order to reduce the probability of a traffic accident occurring when the first vehicle travels at the intersection, when determining the travel track of the first vehicle, the state of a signal lamp corresponding to the target position at the intersection needs to be considered.

Therefore, after the first vehicle acquires the traveling task of the first vehicle, the traveling information of the first vehicle, and the traveling information of the first obstacle vehicle, the first vehicle needs to acquire the traffic light state at the intersection corresponding to the target position.

(2) And determining the running track of the first vehicle through the target neural network model according to the running task and the running information of the first vehicle and the running information of the first obstacle vehicle.

Specifically, the first vehicle may determine the travel track of the first vehicle through the neural network model corresponding to the link, using the travel task of the first vehicle, the travel information of the first vehicle, and the travel information of the first obstacle vehicle as inputs of the neural network model.

Since the target neural network model is trained from the travel tracks of the plurality of second vehicles that have passed through the road section where the target position is located before the current time, the target neural network model has learned the characteristics of the historical travel track, and therefore, when the travel task of the first vehicle, the travel information of the first vehicle, and the travel information of the first obstacle vehicle are input as the target neural network model, the target neural network model can determine the travel track of the first vehicle by the characteristics of the learned historical travel track.

Further, when the road section where the target position of the first vehicle is located is the intersection, at this time, the first vehicle may use the driving task of the first vehicle, the driving information of the first obstacle vehicle, and the signal lamp state at the intersection corresponding to the target position as inputs of the neural network model, and determine the driving track of the first vehicle through the neural network model corresponding to the road section.

The neural network model corresponding to the road section where the target position is located is obtained by the server through training according to the historical driving track of the historical vehicle passing through the road section. The training of the neural network model corresponding to the road segment where the target position is located by the server will be described in detail in the following embodiments, which will not be explained herein.

In addition to providing the method for planning a driving trajectory described in the above embodiments, the embodiments of the present invention also provide a device for planning a driving trajectory, which will be described in the following embodiments.

Fig. 8 is a device 800 for planning a driving trajectory according to an embodiment of the present invention, which is applied to the vehicle shown in fig. 1, and as shown in fig. 8, the device 800 includes a receiving unit 801 and a first determining unit 802:

a receiving unit 801 for performing step 603 in the embodiment shown in fig. 6A or step 708 in the embodiment shown in fig. 7;

a first determining unit 802 for performing step 604 in the embodiment shown in fig. 6A or step 709 in the embodiment shown in fig. 7;

the driving task comprises straight driving, left turning, right turning and turning around, the driving information comprises position information, driving direction and driving speed of a current position, and the first obstacle vehicle is a vehicle, wherein the distance between the first obstacle vehicle and the first vehicle is smaller than a preset distance threshold value.

Optionally, the road segment where the target position is located is an intersection;

the first determining unit 802 is specifically configured to:

Optionally, the target neural network model is a neural network model corresponding to a road segment where the target location is located and a driving task of the first vehicle, and the plurality of second vehicles are vehicles that pass through the road segment where the target location is located before the current time and have the same driving task as the driving task of the first vehicle.

It should be noted that: in the device for planning a driving trajectory according to the above embodiment, when planning a driving trajectory of a first vehicle, only the division of the functional modules is illustrated, and in practical applications, the function distribution may be completed by different functional modules according to needs, that is, the internal structure of the first vehicle may be divided into different functional modules to complete all or part of the functions described above. In addition, the device for planning a driving track provided in the above embodiment and the method for planning a driving track in the above embodiment belong to the same concept, and specific implementation processes thereof are detailed in the method embodiment and are not described herein again.

Fig. 9A is another apparatus 900 for planning a driving trajectory according to an embodiment of the present invention, which is applied to the server shown in fig. 1, and as shown in fig. 9A, the apparatus 900 includes a second determining unit 901 and a sending unit 902:

a second determining unit 901, configured to perform step 601 in the embodiment shown in fig. 6A or step 706 in the embodiment shown in fig. 7;

a sending unit 902, configured to execute step 602 in the embodiment shown in fig. 6A or step 707 in the embodiment shown in fig. 7.

Optionally, referring to fig. 9B, the apparatus 900 further includes a third determining unit 903, a selecting unit 904, and a training unit 905:

a third determining unit 903, configured to perform steps 701 to 703 in the embodiment shown in fig. 7;

a selection unit 904 for performing step 704 in the embodiment shown in fig. 7;

training unit 905, configured to perform step 705 in the embodiment shown in fig. 7.

Optionally, the training unit 905 comprises a first determining subunit and a training subunit:

a first determining subunit operable to determine the travel tasks and the travel information of the N second vehicles, and the travel information of the N second obstacle vehicles;

the N second vehicles correspond to the N running tracks, the N second obstacle vehicles correspond to the N second vehicles one by one, the distance between each second obstacle vehicle and the corresponding second vehicle is smaller than a preset distance threshold value, the running task comprises straight running, left turning, right turning and turning around, and the running information comprises the current position, the running direction and the running speed;

and the training subunit is used for training the initialized neural network model according to the running tasks and the running information of the N second vehicles, the running information of the N second obstacle vehicles and the running tracks of the N second vehicles to obtain the target neural network model.

the training unit 905 further comprises a second determining subunit:

the second determining subunit is used for determining N signal lamp states, the N signal lamp states correspond to the N second vehicles one by one, and each signal lamp state refers to a corresponding signal lamp state at the intersection when the corresponding second vehicle passes through the intersection;

accordingly, the training subunit is specifically configured to:

Optionally, the third determining unit 903 is specifically configured to execute step 702 and step 703 in the embodiment shown in fig. 7.

Optionally, the second determining unit 901 is specifically configured to:

according to the position information of the target position and the driving task of the first vehicle, determining a target neural network model corresponding to the road section where the target position is located and the driving task of the first vehicle from the stored neural network models;

It should be noted that: in the device for planning a driving trajectory according to the above embodiment, when planning a driving trajectory of a first vehicle, only the division of the functional modules is illustrated, and in practical applications, the function distribution may be completed by different functional modules according to needs, that is, the internal structure of the server is divided into different functional modules, so as to complete all or part of the functions described above. In addition, the device for planning a driving track provided in the above embodiment and the method for planning a driving track in the above embodiment belong to the same concept, and specific implementation processes thereof are detailed in the method embodiment and are not described herein again.

In the above embodiments, the implementation may be wholly or partly realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with embodiments of the invention, to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website, computer, server, or data center to another website, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., Digital Versatile Disk (DVD)), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above-mentioned embodiments are provided not to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method of planning a travel path for a first vehicle, the method comprising:

2. The method of claim 1, wherein the road segment where the target location is located is an intersection;

3. The method according to claim 1 or 2, wherein the target neural network model is a neural network model corresponding to both a road segment where the target location is located and a travel task of the first vehicle, and the plurality of second vehicles are vehicles that have passed through the road segment where the target location is located before the current time and have the same travel task as the travel task of the first vehicle.

4. A method for planning a driving track, which is applied to a server, is characterized by comprising the following steps:

5. The method of claim 4, wherein prior to determining the target neural network model from the stored neural network models that corresponds to the segment in which the target location is currently located for the first vehicle, further comprising:

6. The method of claim 5, wherein training the initialized neural network model through the N driving trajectories to obtain the target neural network model comprises:

7. The method of claim 6, wherein the road segment where the target location is located is an intersection;

8. The method of claim 7, wherein determining a score for the travel trajectory of each second vehicle comprises:

9. The method of any one of claims 4 to 8, wherein determining the target neural network model corresponding to the segment where the target position of the first vehicle is currently located from the stored neural network models comprises:

10. An apparatus for planning a driving trajectory for a first vehicle, the apparatus comprising:

a receiving unit, configured to receive a target neural network model sent by a server, where the target neural network model is a neural network model corresponding to a road segment where a target position where the first vehicle is currently located is located, and the target neural network model is obtained by the server through training according to driving tracks of a plurality of second vehicles, where the plurality of second vehicles are vehicles that pass through the road segment where the target position is located before current time;

a first determination unit, configured to determine a travel track of the first vehicle through the target neural network model according to the travel task and the travel information of the first vehicle and the travel information of a first obstacle vehicle;

11. The apparatus of claim 10, wherein the road segment where the target location is located is an intersection;

the first determining unit is specifically configured to:

12. The apparatus according to claim 10 or 11, wherein the target neural network model is a neural network model corresponding to both a road segment where the target location is located and a travel task of the first vehicle, and the plurality of second vehicles are vehicles that have passed through the road segment where the target location is located before a current time and have the same travel task as the travel task of the first vehicle.

13. An apparatus for planning a driving track, applied to a server, the apparatus comprising:

a second determining unit, configured to determine, from stored neural network models, a target neural network model corresponding to a road segment where a target position where a first vehicle is currently located is located, where the target neural network model is obtained through training according to travel tracks of a plurality of second vehicles, where the plurality of second vehicles are vehicles that have passed through the road segment where the target position is located before current time;

and the sending unit is used for sending the target neural network model to the first vehicle so that the first vehicle determines the running track of the first vehicle through the target neural network model.

14. The apparatus of claim 13, further comprising:

the third determining unit is used for determining the driving tracks of all second vehicles passing through the road section where the target position is located in the preset time period and the grade of the driving track of each second vehicle;

the selection unit is used for selecting N driving tracks with scores larger than a preset score from all the obtained driving tracks, wherein N is larger than 1 and smaller than or equal to the total number of the obtained driving tracks;

and the training unit is used for training the initialized neural network model through the N driving tracks to obtain the target neural network model.

15. The apparatus of claim 14, wherein the training unit comprises:

16. The apparatus of claim 15, wherein the road segment where the target location is located is an intersection;

the training unit further comprises:

the second determining subunit is used for determining N signal lamp states, wherein the N signal lamp states correspond to the N second vehicles one by one, and each signal lamp state refers to a corresponding signal lamp state at the intersection when the corresponding second vehicle passes through the intersection;

correspondingly, the training subunit is specifically configured to:

17. The apparatus according to claim 16, wherein the third determining unit is specifically configured to:

18. The apparatus according to any one of claims 13 to 17, wherein the second determining unit is specifically configured to: