CN115320623B - Vehicle trajectory prediction method, apparatus, mobile device, and storage medium - Google Patents

Abstract

Embodiments of the present invention provide a vehicle trajectory prediction method, equipment, mobile device and storage medium. The method includes: when the vehicle enters the intersection, based on the map information of the intersection, determining a first-level exit candidate point for the vehicle's normal driving and a second-level exit candidate point for the vehicle's irregular driving, predicting that the vehicle travels to the first-level exit candidate point, and Multiple predicted trajectories of exit candidate points on the second layer; determine the respective turn width costs of multiple predicted trajectories based on the preset cost function; from the respective turn width costs of multiple predicted trajectories, preferentially select the value of the lowest turn width cost. The corresponding predicted trajectory is used as the trajectory prediction result of the vehicle. The embodiment of the present invention uses the double-layer exit idea to have a good generalization ability for some irregular vehicles, and can also obtain the corresponding trajectory prediction results. At the same time, the cost filter of the historical frame can prevent the predicted trajectory from Frequent fluctuations increase the stability of automatic driving.

Description

Vehicle trajectory prediction method, device, mobile device and storage medium

technical field

The present invention relates to the field of automatic driving, in particular to a vehicle track prediction method, equipment, mobile device and storage medium.

Background technique

With the increasing application of autonomous driving, autonomous driving technology is applied to more and more fields, such as automatic express delivery vehicles equipped with automatic driving systems, automatic floor washing vehicles, etc. The entire automatic driving system mainly includes modules such as positioning-perception-prediction-planning, among which the main function of the prediction module is to predict the trajectory of obstacles in the future based on the perceived obstacle information in the surrounding environment, so as to provide future planning for the own vehicle. Provide some reference. The intersection scene is a typical scene that the prediction module needs to solve, especially the stability and rationality of the vehicle's predicted trajectory in the intersection environment plays a key role in how the self-vehicle makes decisions.

The trajectory prediction of the existing technology in the intersection environment is basically to predict the vehicle on the basis of high-precision maps, and combine the intersection lane information and historical trajectory point information to generate a trajectory after a few seconds. There are two types of prediction methods:

1. Adopt some regularization methods, use the matching relationship between vehicles and roads at the intersection, and some empirical assumptions to establish a regularization method, so as to predict the future trajectory of the vehicle.

2. Based on the prediction method of deep learning, a large number of vehicle driving scenes at intersections are collected as training samples, and then the parameters in the neural network are optimized and adjusted using the idea of deep learning, thereby generating a network model to realize the output of the predicted trajectory.

In the process of realizing the present invention, the inventors have found that there are at least the following problems in the related art:

1. Most of the rule-based prediction methods rely on high-precision maps and normal empirical assumptions to design ideas. Therefore, for some vehicles that do not completely follow the lane lines, the rule-based prediction methods do not have a good generality. ability.

2. The prediction method based on deep learning needs to complete a large amount of data labeling and pre-training work in the early stage, and the final output trajectory lacks certain interpretability.

At the same time, the above two methods do not have a certain correlation test for the predicted trajectory between the front and back frames, which will have an adverse effect on the decision-making planning module at the back end. For example, the vehicle trajectory predicted in the previous frame is to drive to exit A, and the vehicle trajectory predicted in the next frame is to drive to exit B. The frequent jumps of this trajectory will cause the back-end decision-making layer to perform completely different decisions on the self-vehicle. Decision-making logic eventually leads to the instability of automatic driving.

Contents of the invention

In order to at least solve the problem that the prediction method does not have good generalization ability and the obtained predicted trajectory will jump frequently, making the automatic driving unstable when the vehicle is not driving according to the standard and does not completely follow the lane line in the prior art.

In a first aspect, an embodiment of the present invention provides a vehicle trajectory prediction method, including:

When a vehicle enters an intersection, determine a first-level exit candidate point for the vehicle's regulated driving based on the map information of the intersection, and establish a second-level exit candidate point for the vehicle's irregular driving based on the first-level exit candidate point , predicting multiple predicted trajectories of the vehicle traveling to the first layer of exit candidate points and the second layer of exit candidate points;

determining the respective turn amplitude costs of the plurality of predicted trajectories based on a preset cost function, wherein the cost function is related to a position change during driving of the vehicle;

From the respective turning width costs of the plurality of predicted trajectories, the predicted trajectory corresponding to the turning width cost with the lowest numerical value is preferentially selected as the trajectory prediction result of the vehicle.

In a second aspect, an embodiment of the present invention provides a vehicle trajectory prediction system, including:

A candidate point prediction module, configured to determine, based on the map information of the intersection, a first-level exit candidate point for the vehicle's regulated driving when the vehicle enters the intersection, and establish the irregular driving of the vehicle based on the first-level exit candidate point The second layer of exit candidate points, predicting multiple predicted trajectories of the vehicle traveling to the first layer of exit candidate points and the second layer of exit candidate points;

A cost determination module, configured to determine the respective turning amplitude costs of the plurality of predicted trajectories based on a preset cost function, wherein the cost function is related to the position change of the vehicle during driving;

The trajectory prediction module is configured to preferentially select the predicted trajectory corresponding to the lowest valued turning width cost from the respective turning width costs of the plurality of predicted trajectories as the trajectory prediction result of the vehicle.

In a third aspect, an electronic device is provided, which includes: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, The instructions are executed by the at least one processor, so that the at least one processor can execute the steps of the vehicle trajectory prediction method in any embodiment of the present invention.

In a fourth aspect, an embodiment of the present invention provides a mobile device, including a body and the electronic device according to any embodiment of the present invention installed on the body.

In a fifth aspect, an embodiment of the present invention provides a storage medium on which a computer program is stored, wherein when the program is executed by a processor, the steps of the vehicle trajectory prediction method in any embodiment of the present invention are implemented.

In a sixth aspect, an embodiment of the present invention further provides a computer program product, which, when running on a computer, causes the computer to execute the vehicle trajectory prediction method described in any one of the embodiments of the present invention.

The beneficial effect of the embodiment of the present invention is that: the use of the double-layer exit idea has good generalization ability for some vehicles with irregular driving, and the corresponding trajectory prediction results can also be obtained. Filtering can prevent frequent fluctuations in the predicted trajectory and increase the stability of automatic driving.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will give a brief introduction to the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is a flow chart of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 2 is a schematic diagram of an intersection map environment provided by a vehicle trajectory prediction method according to an embodiment of the present invention;

Fig. 3 is a schematic diagram of exit candidate points of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 4 is a schematic diagram of the predicted trajectory of each candidate point in a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 5 is a schematic diagram of the intersection exit classification of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 6 is a schematic diagram of the degree of curvature change of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 7 is a schematic diagram of angle change degree of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 8 is a schematic diagram of the distance change degree of a vehicle trajectory prediction method provided by an embodiment of the present invention;

Fig. 9 is a schematic structural diagram of a vehicle trajectory prediction execution device provided by an embodiment of the present invention;

Fig. 10 is a schematic structural diagram of an embodiment of an electronic device for vehicle trajectory prediction provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

Those skilled in the art know that the embodiments of the present application may be realized as a system, device, device, method or computer program product. Therefore, the present disclosure may be specifically implemented in the form of complete hardware, complete software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

For ease of understanding, the technical terms involved in this application are explained below:

The "mobile device" referred to in this application includes but is not limited to L0-L5 autonomous driving technology level vehicles formulated by the Society of Automotive Engineers International (SAE International) or China's national standard "Automotive Driving Automation Classification".

In some embodiments, the mobile device may be a vehicular device or a robotic device with various functions as follows:

(1) Passenger functions, such as family cars, buses, etc.;

(2) Loading function, such as ordinary trucks, box trucks, drop trailers, closed trucks, tank trucks, flatbed trucks, container trucks, dump trucks, special structure trucks, etc.;

(3) Tool functions, such as logistics distribution vehicles, automatic guided transport vehicles AGV, patrol vehicles, cranes, cranes, excavators, bulldozers, forklifts, road rollers, loaders, off-road engineering vehicles, armored engineering vehicles, sewage treatment vehicles, Sanitation vehicles, vacuum cleaners, floor washing vehicles, sprinklers, sweeping robots, food delivery robots, shopping guide robots, lawn mowers, golf carts, etc.;

(4) Entertainment functions, such as recreational vehicles, playground automatic driving devices, balance vehicles, etc.;

(5) Special rescue functions, such as fire engines, ambulances, electric emergency repair vehicles, engineering emergency vehicles, etc.

As shown in Figure 1, it is a flowchart of a vehicle trajectory prediction method provided by an embodiment of the present invention, including the following steps:

S11: When the vehicle enters an intersection, determine a first-level exit candidate point where the vehicle is regulated based on the map information of the intersection, and establish a second-level exit where the vehicle is irregularly driven based on the first-level exit candidate point Candidate points, predicting multiple predicted trajectories for the vehicle to travel to the first layer of exit candidate points and the second layer of exit candidate points;

S12: Determine the respective turning width costs of the plurality of predicted trajectories based on a preset cost function, wherein the cost function is related to the position change of the vehicle during driving;

S13: From the respective turning width costs of the multiple predicted trajectories, preferentially select the predicted trajectory corresponding to the turning width cost with the lowest numerical value as the trajectory prediction result of the vehicle.

In this embodiment, this method can be applied to various types of automatic driving devices, such as automatic express delivery vehicles in communities, automatic food distribution vehicles in restaurants, and automatic floor washing vehicles in industrial parks. Taking the automatic floor washing vehicle in the industrial park as an example, the automatic floor washing vehicle stores a high-definition map of the industrial park, and can accurately obtain the environmental conditions of each intersection in the industrial park through the data stored in the high-definition map. In this method, the trajectory prediction of the vehicle is based on determining the exit of the intersection that the vehicle will drive to, as shown in FIG. 2 , which is the intersection map environment.

For step S11, take an automatic floor washing vehicle in an industrial park as an example (the automatic driving device may also be other devices, which are not limited here). When the automatic floor washing vehicle drives to the intersection in Figure 2, it is first necessary to determine the possible exits of the automatic floor washing vehicle, and obtain the number and specific location of the exits based on the intersection information in the existing high-precision map information, as a vehicle specification driving The first layer of exit candidate points.

However, considering the real intersection environment, the automatic floor washing vehicle does not necessarily drive completely according to the lane line, or it does not target the exit point, but may target the position after the exit. In order to be compatible with the above-mentioned special circumstances, on the basis of the determined position of the exit candidate point of the first layer, it is shifted to the outside by a certain distance d, where the value of the distance d is configurable, and then the second layer as shown in Figure 3 is established. Export candidate points.

After calculating the positions of the double-layer exit points, it is determined that all the candidate points in the first-level exit candidate points and the second-level exit candidate points are the possible arrival positions of the automatic washing machine, so that the Bezier curve can be used To generate the predicted trajectory of the automatic floor washing vehicle to each candidate point as shown in Figure 4, other methods can also be used, such as dubins curve, polynomial fitting and other methods, which are not limited here.

For step S12, after obtaining various possible predicted trajectories, it is necessary to calculate the turning range cost of each trajectory, and finally compare and select the trajectory with the lowest cost value, considering that the intersection is a potentially dangerous scene , during the turning process at the intersection, the turning range at the intersection should be reduced as much as possible. The trajectory with the lowest cost value is equivalent to the relatively small driving range of the trajectory at the intersection, and the smaller the potential safety hazard. As the final prediction track.

Specifically, the cost function of multiple dimensions can be determined through the position change of the automatic floor washing vehicle. For example, by determining the specific position of the automatic washing vehicle in the map during driving, the lane line where the automatic washing vehicle is currently located can be determined in real time; by using the coordinate changes in the position of the automatic washing vehicle, the automatic washing vehicle can be determined Curvature, declination, distance, etc. during driving.

As an implementation manner, the position change during the driving of the vehicle includes a change in the position of the lane line where the vehicle is located during the driving, and the determination of the respective turning amplitude costs of the plurality of predicted trajectories based on the preset cost function includes:

The lane matching cost in the cost function is determined by matching the current lane line position of the vehicle with the exit candidate point, wherein the smaller the lane matching cost is, the smaller the turning width cost is.

In this embodiment, the lane where the automatic floor washing vehicle is currently located is obtained, denoted as L. Match the lane where the exit position is located with the current lane L, and divide the matching relationship into three categories: the exit lane and L belong to the same lane or have a front-rear relationship, which is recorded as type A; the exit lane has an adjacent relationship with L or has a relationship with L The lower-level lanes of L have adjacent relations, which are recorded as Class B; if there is no relationship, they are recorded as Class C. As shown in Figure 5, according to different matching relationships, set the cost Cost_lane as follows

Among them, c_A, c_B, and c_C are configurable items. The smaller the lane matching cost value determined according to different categories, the smaller the turning width cost.

As an implementation manner, the position change during driving of the vehicle includes a curvature change of the vehicle during driving, and the determining the respective turning amplitude costs of the plurality of predicted trajectories based on a preset cost function includes:

determining the historical curvature of the vehicle using the intersection location and the vehicle's current location;

determining the predicted curvature of the vehicle by the current position of the vehicle and the exit position of the predicted trajectory;

A curvature change cost in the cost function is determined based on the historical curvature and the predicted curvature, wherein the smaller the curvature change cost is, the smaller the turn width cost is.

In this embodiment, as shown in FIG. 6 , the recorded position of the automatic floor washing vehicle just entering the intersection is point A, the current position of the automatic floor washing vehicle is point B, and the recorded position of the exit of the intersection is point C. Calculate the maximum curvature c_his of the historical trajectory of the automatic scrubber (that is, the trajectory of the automatic scrubber from point A to point B), and calculate the maximum curvature c_his of the future predicted trajectory of the automatic scrubber (that is, the trajectory of the automatic scrubber from point B to point C). curvature c_fur.

The change degree cost of curvature is generated by the change relationship between historical curvature and predicted curvature. The setting rule is that the smaller the predicted curvature is, the smaller the cost of the turning range is, based on which the following cost function is established (k is a configurable coefficient in the following formula):

As an implementation manner, the position change during driving of the vehicle includes a change in the deflection angle of the vehicle during driving, and the determining the respective turning amplitude costs of the plurality of predicted trajectories based on the preset cost function includes:

Determine the first orientation of the vehicle at the intersection, the second orientation of the current position, and the third orientation of the exit position;

Using the degree of deviation of the first orientation and the second orientation relative to the third orientation respectively to determine a deflection angle change cost in the cost function, wherein the smaller the deflection angle change cost, the less the turning The smaller the magnitude cost.

In this embodiment, considering that the turning range of the automatic floor washing vehicle at the intersection is also related to the variation of the deflection angle, as shown in FIG. The current position of the ground car is B, its orientation is angleB, the position of the recorded exit is C, and its orientation is angleC.

Calculate the deviation degree of angleA and angleB relative to angleC to generate the cost of declination change (k in the following formula is a configurable coefficient):

As an implementation manner, the position change during the driving of the vehicle includes the trajectory change of the vehicle during the driving, and the determination of the respective turning amplitude costs of the plurality of predicted trajectories based on the preset cost function includes:

Determine the historical trajectory length of the vehicle traveling from the intersection to the current location, and the predicted trajectory length of the vehicle traveling from the current location to the exit location;

A distance change cost in the cost function is generated based on the historical trajectory length and the predicted trajectory length, wherein the smaller the distance change cost is, the smaller the turn width cost is. In this embodiment, it is considered that the turning range of the automatic floor washing vehicle at the intersection is also related to the distance change, as shown in FIG. 8 .

Record the position of the automatic floor washing vehicle just entering the intersection as point A, the current position of the automatic floor washing vehicle as point B, and record the position of the exit of the intersection as point C. The length of the historical trajectory is denoted as L_AB, and the length of the predicted trajectory is denoted as L_BC.

The distance change cost is generated according to the two lengths L_AB and L_BC (k in the following formula is a configurable parameter):

Calculate the lane matching cost, curvature change cost, deviation angle change cost, and distance change through the preset cost function including lane matching cost function, curvature change cost function, declination angle change cost function, and distance change cost function in the above steps The turning range cost of the cost can be determined by the following formula:

Among them, Cost_j represents the cost function of a certain exit point j.

For step S13, after the calculation of the above costs, the final turning width cost of each exit in the first layer of exit candidate points and the second layer of exit candidate points can be obtained, and then select the exit with the smallest turning width cost value from all exits , it is used as the optimal prediction exit, and the generated prediction trajectory is used as the optimal prediction trajectory.

As an implementation manner, during the driving process of the vehicle according to the trajectory prediction result, the method further includes:

Determine the respective turn amplitude costs of multiple predicted trajectories of the current frame;

Perform filtering processing on the respective turning amplitude costs of the multiple predicted trajectories in the existing history and the respective turning amplitude costs of the multiple predicted trajectories in the current frame to obtain an optimized turning amplitude cost;

From the respective turning amplitude costs of the optimized plurality of predicted trajectories, the predicted trajectory corresponding to the lowest numerical value of the turning amplitude cost is preferentially selected as the trajectory prediction result of the vehicle, so that the vehicle can continue to follow the trajectory prediction result. Drive steadily.

In this embodiment, after the design of the above-mentioned exit and the calculation of the turning width cost, the exit and trajectory predicted in the intersection environment can be obtained, but for the prediction module, the stability and continuity of the predicted trajectory are also important. index. That is to say, the automatic floor washing vehicle does not always drive according to the predicted trajectory determined at the intersection, but needs to determine the current latest predicted trajectory in real time. However, it will appear that the vehicle trajectory predicted in the previous frame is to drive to exit A, and the vehicle trajectory predicted in the next frame is to drive to exit B. The frequent jumps of this trajectory will cause the back-end decision-making layer to execute completely differently for the own vehicle. The decision-making logic eventually leads to the unstable situation of automatic driving.

In order to avoid this situation, when the vehicle is traveling according to the trajectory, this method compares and filters the current predicted exit and the previous predicted exit (other filtering methods can also be used, which are not limited here) to obtain the final optimization subsequent predictions. Obtain the prediction results of the historical n frames of each exit point and perform coupling processing with the prediction results of the current frame to obtain the predicted exit cost results of n+1 frames including the current frame, as follows:

Record the cost value of a certain exit point j in the current frame as Cost_j, and record the cost value of a certain exit point j in the previous i frame as Cost_j_i. In order to filter the cost value of the current frame in combination with the cost value of the historical frame, set the weight coefficient under the current frame to w_cur (can be arbitrarily configured between 0 and 1, here is set to 0.5), and the weight coefficient of the historical frame is w_his , the relationship between the two is:

Among them, n is the number of historical frames.

Using the cost value of the historical frame and the cost value of the current frame, a new cost is calculated by using the filtering algorithm as the cost after the optimization of the current frame, which is recorded as NewCost_j, then the new cost value of a certain exit point j is calculated as:

Then compare the cost values of all the exits after the update, and select the exit with the smallest cost value as the final predicted exit, and determine the real-time vehicle predicted trajectory results through the predicted exit.

It can be seen from this implementation that the use of the double-layer exit idea has a good generalization ability for some irregular vehicles, and the corresponding trajectory prediction results can also be obtained. At the same time, the cost filtering of historical frames is combined in the process of automatic driving. It can prevent frequent fluctuations in the predicted trajectory and increase the stability of automatic driving.

FIG. 9 is a schematic structural diagram of a vehicle trajectory prediction execution device provided by an embodiment of the present invention. The system can execute the vehicle trajectory prediction method described in any of the above embodiments, and is configured in a terminal.

A vehicle trajectory prediction system 10 provided in this embodiment includes: a screen configuration acquisition program module 11 , a layout configuration adjustment program module 12 and a screen adaptation program module 13 .

Wherein, the candidate point prediction module 11 is used to determine the first layer of exit candidate points for the vehicle to standardly drive based on the map information of the intersection when the vehicle enters the intersection, and establish the vehicle's non-exit candidate points based on the first layer of exit candidate points. Standardize the second-level exit candidate points for driving, and predict multiple predicted trajectories for the vehicle to travel to the first-level exit candidate points and the second-level exit candidate points; the cost determination module 12 is used to The function determines the respective turning amplitude costs of the plurality of predicted trajectories, wherein the cost function is related to the position change during the driving of the vehicle; the trajectory prediction module 13 is used to obtain from the respective turning amplitude costs of the plurality of predicted trajectories , the predicted trajectory corresponding to the lowest turning amplitude cost is preferentially selected as the trajectory prediction result of the vehicle.

Further, the trajectory prediction module is also used to: determine the respective turning amplitude costs of multiple predicted trajectories of the current frame;

Perform filtering processing on the respective turning amplitude costs of the multiple predicted trajectories in the existing history and the respective turning amplitude costs of the multiple predicted trajectories in the current frame to obtain an optimized turning amplitude cost;

From the respective turning amplitude costs of the optimized plurality of predicted trajectories, the predicted trajectory corresponding to the lowest numerical value of the turning amplitude cost is preferentially selected as the trajectory prediction result of the vehicle, so that the vehicle can continue to follow the trajectory prediction result. Drive steadily.

Further, the cost determination module is used for:

The lane matching cost in the cost function is determined by matching the current lane line position of the vehicle with the exit candidate point, wherein the smaller the lane matching cost is, the smaller the turning width cost is.

Further, the cost determination module is used for:

determining the historical curvature of the vehicle using the intersection location and the vehicle's current location;

determining the predicted curvature of the vehicle by the current position of the vehicle and the exit position of the predicted trajectory;

A curvature change cost in the cost function is determined based on the historical curvature and the predicted curvature, wherein the smaller the curvature change cost is, the smaller the turn width cost is.

Further, the cost determination module is used for:

Determine the first orientation of the vehicle at the intersection, the second orientation of the current position, and the third orientation of the exit position;

Using the degree of deviation of the first orientation and the second orientation relative to the third orientation respectively to determine a deflection angle change cost in the cost function, wherein the smaller the deflection angle change cost, the less the turning The smaller the magnitude cost.

Further, the cost determination module is used for:

Determine the historical trajectory length of the vehicle traveling from the intersection to the current location, and the predicted trajectory length of the vehicle traveling from the current location to the exit location;

A distance change cost in the cost function is generated based on the historical trajectory length and the predicted trajectory length, wherein the smaller the distance change cost is, the smaller the turn width cost is.

The embodiment of the present invention also provides a non-volatile computer storage medium, the computer storage medium stores computer-executable instructions, and the computer-executable instructions can execute the vehicle trajectory prediction method in any of the above-mentioned method embodiments;

As an implementation mode, the non-volatile computer storage medium of the present invention stores computer-executable instructions, and the computer-executable instructions are set to:

When a vehicle enters an intersection, determine a first-level exit candidate point for the vehicle's regulated driving based on the map information of the intersection, and establish a second-level exit candidate point for the vehicle's irregular driving based on the first-level exit candidate point , predicting multiple predicted trajectories of the vehicle traveling to the first layer of exit candidate points and the second layer of exit candidate points;

determining the respective turn amplitude costs of the plurality of predicted trajectories based on a preset cost function, wherein the cost function is related to a position change during driving of the vehicle;

From the respective turning width costs of the plurality of predicted trajectories, the predicted trajectory corresponding to the turning width cost with the lowest numerical value is preferentially selected as the trajectory prediction result of the vehicle.

As a non-volatile computer-readable storage medium, it can be used to store non-volatile software programs, non-volatile computer-executable programs and modules, such as program instructions/modules corresponding to the methods in the embodiments of the present invention. One or more program instructions are stored in a non-volatile computer-readable storage medium, and when executed by a processor, execute the vehicle trajectory prediction method in any of the above method embodiments.

An embodiment of the present invention also provides an electronic device, which includes: at least one processor, and a memory connected in communication with the at least one processor, wherein the memory stores instructions that can be executed by the at least one processor , the instructions are executed by the at least one processor, so that the at least one processor can execute and apply a vehicle trajectory prediction method.

In some embodiments, an embodiment of the present invention further provides a mobile device, including a body and the electronic device according to any one of the preceding embodiments installed on the body. Among them, the mobile device may be an unmanned vehicle, such as an unmanned sweeper, an unmanned floor scrubber, an unmanned logistics vehicle, an unmanned passenger vehicle, an unmanned sanitation vehicle, and an unmanned minibus. / Buses, trucks, mining carts, etc., can also be robots, etc.

In some embodiments, the embodiments of the present invention also provide a computer program product, which, when the computer program product is run on a computer, causes the computer to execute the vehicle trajectory prediction method described in any one of the embodiments of the present invention .

Fig. 10 is a schematic diagram of the hardware structure of the electronic device of the vehicle trajectory prediction method provided by another embodiment of the present application. As shown in Fig. 10, the device includes:

One or more processors 1010 and memory 1020, one processor 1010 is taken as an example in FIG. 10 . The equipment of the vehicle trajectory prediction method may further include: an input device 1030 and an output device 1040 .

The processor 1010, the memory 1020, the input device 1030, and the output device 1040 may be connected via a bus or in other ways, and connection via a bus is taken as an example in FIG. 10 .

The memory 1020, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs and modules, such as the program corresponding to the vehicle trajectory prediction method in the embodiment of the present application directive/module. The processor 1010 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 1020, that is, implements the vehicle trajectory prediction method in the above method embodiment.

The memory 1020 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application program required by at least one function; the data storage area may store data and the like. In addition, the memory 1020 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 1020 may optionally include memory located remotely from the processor 1010, and these remote memories may be connected to the mobile device through a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The input device 1030 can receive input numbers or character information. The output device 1040 may include a display device such as a display screen.

The one or more modules are stored in the memory 1020, and when executed by the one or more processors 1010, execute the vehicle trajectory prediction method in any method embodiment above.

The above-mentioned products can execute the method provided by the embodiment of the present application, and have corresponding functional modules and beneficial effects for executing the method. For technical details not described in detail in this embodiment, refer to the method provided in the embodiment of this application.

The non-volatile computer-readable storage medium may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function; the data storage area may store the data etc. In addition, the non-volatile computer-readable storage medium may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the non-transitory computer-readable storage medium optionally includes memory located remotely from the processor, which remote memory may be connected to the device via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

An embodiment of the present invention also provides an electronic device, which includes: at least one processor, and a memory connected in communication with the at least one processor, wherein the memory stores instructions that can be executed by the at least one processor , the instructions are executed by the at least one processor, so that the at least one processor can execute the steps of the vehicle trajectory prediction method in any embodiment of the present invention.

The electronic device in the embodiment of the present application may be an automatic driving domain controller applied to a mobile device, which communicates with various sensors (such as laser radar, camera, millimeter wave radar, ultrasonic radar, inertial Measuring unit IMU, wheel speedometer, etc.) communication connection, through these sensors to obtain environmental perception data and vehicle body speed information, and extract obstacle information and generate mobile device position information according to the acquired environment perception data and vehicle body speed information, and according to the position information and obstacle information for path planning.

The electronic devices in the embodiments of this application can also exist in the following forms, including but not limited to:

(1) Mobile communication equipment: This type of equipment is characterized by mobile communication functions, and its main goal is to provide voice and data communication. Such terminals include: smart phones, multimedia phones, feature phones, and low-end phones.

(2) Ultra-mobile personal computer equipment: This type of equipment belongs to the category of personal computers, has computing and processing functions, and generally has the characteristics of mobile Internet access. Such terminals include: PDA, MID and UMPC equipment, etc., such as tablet computers.

(3) Portable entertainment equipment: This type of equipment can display and play multimedia content. Such devices include: audio and video players, handheld game consoles, e-books, as well as smart toys and portable car navigation devices.

(4) Other mobile devices with data processing functions.

In this document, relational terms such as first and second etc. are used only to distinguish one entity or operation from another without necessarily requiring or implying any such relationship between these entities or operations. Actual relationship or sequence. Moreover, the terms "comprising" and "comprising" not only include those elements, but also include other elements not explicitly listed, or also include elements inherent in such a process, method, article, or device. Without further limitations, an element defined by the statement "comprising..." does not exclude the presence of additional same elements in the process, method, article or device comprising said element.

The device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in One place, or it can be distributed to multiple network elements. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. It can be understood and implemented by those skilled in the art without any creative efforts.

Through the above description of the implementations, those skilled in the art can clearly understand that each implementation can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the prior art can be embodied in the form of software products, and the computer software products can be stored in computer-readable storage media, such as ROM/RAM, magnetic CD, CD, etc., including several instructions to make a computer device (which may be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (10)

1. A vehicle trajectory prediction method, comprising:

when a vehicle enters an intersection, determining a first-layer exit candidate point for the vehicle to normally travel based on map information of the intersection, establishing a second-layer exit candidate point for the vehicle to not normally travel based on the first-layer exit candidate point, and predicting a plurality of predicted tracks of the vehicle traveling to the first-layer exit candidate point and the second-layer exit candidate point;

determining respective turn amplitude costs of the plurality of predicted trajectories based on a preset cost function, wherein the cost function is related to position change in the vehicle in the driving process;

and preferentially selecting the predicted track corresponding to the turn amplitude cost with the lowest numerical value from the turn amplitude costs of the plurality of predicted tracks as the track prediction result of the vehicle.

2. The method of claim 1, wherein during travel of the vehicle in accordance with the trajectory prediction, the method further comprises:

determining the turn amplitude cost of each of a plurality of predicted tracks of the current frame;

filtering the turn amplitude cost of each of the plurality of predicted tracks with the history and the turn amplitude cost of each of the plurality of predicted tracks of the current frame to obtain an optimized turn amplitude cost;

and preferentially selecting the predicted track corresponding to the turn amplitude cost with the lowest numerical value from the turn amplitude costs of the optimized predicted tracks as the track prediction result of the vehicle, so that the vehicle continuously and stably runs according to the track prediction result.

3. The method of claim 1, wherein the change in the position of the vehicle while traveling comprises a change in a lane position of the vehicle while traveling, and wherein determining the turn magnitude cost for each of the plurality of predicted trajectories based on a preset cost function comprises:

and matching the current lane line position of the vehicle with an exit candidate point to determine lane matching cost in the cost function, wherein the smaller the lane matching cost is, the smaller the turn amplitude cost is.

4. The method of claim 1, wherein the change in position of the vehicle while traveling comprises a change in curvature of the vehicle while traveling, and wherein determining the turn magnitude cost for each of the plurality of predicted trajectories based on a preset cost function comprises:

determining a historical curvature of the vehicle using the intersection location and the current location of the vehicle;

determining a predicted curvature of the vehicle from a current position of the vehicle and an exit position of a predicted trajectory;

determining a curvature change cost in the cost function based on the historical curvature and the predicted curvature, wherein the smaller the curvature change cost is, the smaller the turn amplitude cost is.

5. The method of claim 1, wherein the change in position while the vehicle is traveling comprises a change in a yaw angle of the vehicle while traveling, and wherein determining the respective turn magnitude costs for the plurality of predicted trajectories based on a preset cost function comprises:

determining a first orientation of the vehicle at the intersection, a second orientation of the current position, and a third orientation of the exit position;

and determining a yaw angle change cost in the cost function by respectively utilizing the deviation degrees of the first orientation and the second orientation relative to the third orientation, wherein the smaller the yaw angle change cost is, the smaller the turn amplitude cost is.

6. The method of claim 1, wherein the change in position while the vehicle is traveling comprises a change in trajectory of the vehicle while traveling, and wherein determining the respective turn magnitude costs for the plurality of predicted trajectories based on a preset cost function comprises:

determining a historical track length of the vehicle from the intersection to the current position and a predicted track length of the vehicle from the current position to the exit position;

and generating a distance change cost in the cost function based on the historical track length and the predicted track length, wherein the smaller the distance change cost is, the smaller the turn amplitude cost is.

7. A vehicle trajectory prediction execution apparatus characterized by comprising:

the candidate point prediction module is used for determining a first-layer exit candidate point for the vehicle to normally run based on map information of an intersection when the vehicle enters the intersection, establishing a second-layer exit candidate point for the vehicle to not normally run based on the first-layer exit candidate point, and predicting a plurality of predicted tracks of the vehicle running to the first-layer exit candidate point and the second-layer exit candidate point;

the cost determination module is used for determining the turn amplitude cost of each of the plurality of predicted tracks based on a preset cost function, wherein the cost function is related to the position change of the vehicle in the running process;

and the track prediction module is used for preferentially selecting the predicted track corresponding to the turn amplitude cost with the lowest numerical value from the turn amplitude costs of the predicted tracks as the track prediction result of the vehicle.

8. An electronic device, comprising: at least one processor, and a memory communicatively coupled to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the steps of the method of any of claims 1-6.

9. A mobile device, comprising: the electronic device of claim 8.

10. A storage medium on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 6.

Images