An autonomous driving vehicle motion planning method, device and storage medium based on trajectory re-prediction

Abstract

The present application discloses a method, device and storage medium for motion planning of an autonomous driving vehicle based on trajectory re-prediction, which are used to improve the safety, reliability and explainability of the planned path. The method for motion planning of an autonomous driving vehicle based on trajectory re-prediction disclosed in the present application includes: determining preset features and pre-processing the preset features; determining a deep neural network, which is used to predict the motion trajectories of other vehicles in the future time domain; establishing a multi-dimensional space occupancy grid network diagram in the future time domain; checking the trajectory pairs for conflicting behaviors, screening and recording relevant information of trajectory pairs with conflicting behaviors; re-predicting the screened trajectory pairs with conflicting behaviors; and planning the vehicle motion path according to the re-prediction results. The present application also provides a device and storage medium for motion planning of an autonomous driving vehicle based on trajectory re-prediction.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method, device, and storage medium for autonomous driving vehicle motion planning based on trajectory re-prediction. Background Art

[0002] The motion planning module of an autonomous driving system uses perceived obstacle information, road, and roadside information to generate a safe driving path for the vehicle, subject to certain constraints. Existing motion planning methods can be divided into those based on instantaneous states and those based on long-term prediction. However, instantaneous state-based planning methods only consider the current surrounding traffic environment and then search for continuous path points within a desired range that meet the constraints. They fail to account for the dynamic changes of traffic participants, resulting in less safe paths. Long-term prediction-based planning methods consider future changes in the surrounding traffic environment and select safer paths by calculating collision probabilities or modeling them in spatiotemporal dimensions. However, these methods require high prediction accuracy, and the accuracy of trajectory predictions for other traffic participants directly impacts the reliability and safety of the planned path. Therefore, improving trajectory prediction accuracy while enhancing safety, reliability, and interpretability are pressing technical challenges. Summary of the Invention

[0003] In response to the above technical problems, the embodiments of the present application provide a method, device and storage medium for autonomous driving vehicle motion planning based on trajectory re-prediction, so as to improve the accuracy of trajectory prediction and enhance safety, reliability and explainability.

[0004] In a first aspect, an embodiment of the present application provides a motion planning method for an autonomous driving vehicle based on trajectory re-prediction, comprising:

[0005] Determining preset features and preprocessing the preset features;

[0006] Determine a deep neural network for predicting the future time domain Δt fut The movement trajectories of other vehicles in the

[0007] Establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within;

[0008] Check the trajectory pairs for conflicting behaviors, filter and record relevant information of trajectory pairs with conflicting behaviors;

[0009] Re-predict the selected trajectory pairs with conflicting behaviors;

[0010] Plan the vehicle's motion path based on the re-prediction results.

[0011] Preferably, determining the preset features and preprocessing the preset features includes:

[0012] The preset features include one or a combination of the following:

[0013] Longitudinal position x, lateral position y, velocity v, acceleration a, yaw angle yaw, distance to the vehicle in front d f 、Previous vehicle type T f , road type R;

[0014] The pretreatment includes:

[0015] The preceding vehicle type T f , the road type R is numerically represented and the sum of historical driving information past of all surrounding vehicles is standardized.

[0016] Specifically, the preset features satisfy the following relationship:

[0017]

[0018] Among them, n represents the vehicle number, N represents the total number of surrounding vehicles, represents the longitudinal position, lateral position, speed, acceleration, yaw angle, distance to the preceding vehicle, preceding vehicle type and road type of vehicle numbered n at time t, respectively. past Represents the set historical time, past n represents Δt past Vectorized representation of the driving information of vehicle numbered n within a time period, where t0 represents the current moment and past represents the sum of historical driving information of all surrounding vehicles.

[0019] Preferably, the preset characteristics satisfy the following relationship:

[0020]

[0021] Among them, n represents the vehicle number, N represents the total number of surrounding vehicles, represents the longitudinal position, lateral position, speed, acceleration, yaw angle, distance to the preceding vehicle, preceding vehicle type and road type of vehicle numbered n at time t, respectively. past Represents the set historical time, past n represents Δt past Vectorized representation of the driving information of vehicle numbered n within a time period, where t0 represents the current moment and past represents the sum of historical driving information of all surrounding vehicles.

[0022] The deep neural network is used to predict the future time domain Δt fut The motion trajectories of other vehicles in the include:

[0023] The future motion trajectory of vehicle numbered n is pre n for:

[0024] pre n =[X n ,Y n ]=Model(past n ,Action m )

[0025]

[0026] Action m =[1,2,...,M],M≥1

[0027] Among them, X n ,Y n The vehicle numbered n is in the future time domain Δt fut Vector representation of vertical and horizontal positions, Action m Represents the driving behavior of other cars, M represents the number of preset discrete driving behaviors, m is the number of preset discrete driving behaviors, Model() represents different Action m The corresponding prediction model.

[0028] Preferably, the establishing of the multi-dimensional space occupancy grid network diagram in the future time domain includes:

[0029] Mapping the motion trajectories of the other vehicles into a multi-dimensional space occupancy grid map MSOGM;

[0030] Perform Gaussian noise processing on the mapped trajectory points.

[0031] Preferably, the step of checking the trajectory pairs for conflicting behaviors and screening and recording relevant information of the trajectory pairs with conflicting behaviors includes:

[0032] Calculate the lateral and longitudinal safety distances that two vehicles need to maintain;

[0033] Among them, the rear car C r Should be with the front car C f Maintain longitudinal safety distance d 0,lon for:

[0034]

[0035] Among them, v r Represents the longitudinal velocity of the following vehicle at the initial moment, v f represents the longitudinal velocity of the preceding vehicle at the initial moment, ρ represents the reaction time, and a max,accel,lon Represents the maximum longitudinal forward acceleration of the vehicle, a min,brake,lonRepresents the minimum longitudinal braking acceleration of the vehicle, a max,brake,lon Represents the maximum longitudinal braking acceleration of the vehicle;

[0036] Left car C l Should be with the right car C r Minimum lateral safety distance d 0,lat for:

[0037]

[0038] Among them, μ represents the minimum lateral distance between the two vehicles to maintain safety, v1 represents the initial lateral speed of the left vehicle, v2 represents the initial lateral speed of the right vehicle, and v 1,ρ represents the lateral velocity of the left vehicle after time ρ and v 1,ρ =v1+ρa max,accel,lat , a max,accel,lat Represents the maximum lateral acceleration of the vehicle, v 2,ρ represents the lateral velocity of the right vehicle after time ρ and v 2,ρ =v2-ρa max,accle,lat , a min,brake,lat Represents the minimum lateral acceleration of the vehicle.

[0039] Preferably, the checking of conflicting behaviors on the trajectory pair includes:

[0040] Conflict i,j The judgment basis is:

[0041]

[0042] When Conflict i,j A value of 1 indicates that the predicted trajectory pair has conflicting behavior;

[0043] in, They represent the longitudinal and lateral coordinates of the vehicle numbered i at time t, They represent the longitudinal and lateral coordinates of the vehicle numbered j at time t respectively; k1, k2 represent the safety distance coefficients and k1, k2∈R, where R represents a real number.

[0044] Preferably, re-predicting the screened trajectory pairs with conflicting behaviors includes:

[0045] For vehicles with conflicting behaviors, the corresponding driving behaviors are output in combination with the RSS principle:

[0046]

[0047] Action reaRepresents the corrected driving behavior of a vehicle with conflicting behavior after RSS analysis. Represents the action corresponding to vehicle number n rea The trajectory of the next heavy forecast;

[0048] The re-predicted trajectory is re-projected into the multi-dimensional space occupancy grid map MSOGM to obtain a new MSOGM.

[0049] As a preferred example, the autonomous driving vehicle motion planning method of the present invention may further include:

[0050] If the calculation time of the re-prediction exceeds the preset threshold, the prediction is re-performed.

[0051] Preferably, planning the vehicle motion path according to the re-prediction result includes:

[0052] Planning a vehicle motion path according to the new MSOGM, a preset objective function and constraints;

[0053] Wherein, the new MSOGM is:

[0054] MSOGM=[X ts ,Y ts ,T ts ]

[0055] X ts =[x k ],k∈[x min ,x max ]

[0056] Y ts =[y l ],l∈[y miin ,y max ]

[0057] T ts =[t o ],o∈[t0,t0+Δt fut ]

[0058] Among them, X ts ,Y ts ,T ts The vector representation of the vertical coordinate dimension, horizontal coordinate dimension and prediction time domain dimension of the MSOGM is represented in sequence, x k is the vertical position, y l is the horizontal position, t o is the coordinate scale on the time dimension, x min ,x max Represents the minimum, maximum, and y dimensions respectively. min ,y maxRepresent the minimum and maximum values of the y dimension, t0, t0+Δt respectively pre Represent the minimum and maximum values of the t dimension respectively;

[0059] The preset objective function and constraints are:

[0060]

[0061] path risk Represents the comprehensive risk value of each path, represents the longitudinal distance between the vehicle and the nearest neighbor vehicle j at time t, represents the lateral distance between the vehicle and the nearest neighbor vehicle j at time t, Respectively represent the longitudinal position and lateral position of the vehicle at time t, They represent the longitudinal position and lateral position of the nearest neighbor vehicle j at time t, respectively. k and q represent the distance risk coefficients, and k, q∈R, where R is a real number.

[0062] The motion planning method for autonomous vehicles based on trajectory re-prediction provided by the present invention involves, first, extracting relevant features from the perception module and performing pre-processing. Second, a properly designed deep neural network is designed to output the driving behavior of surrounding vehicles, with corresponding confidence levels included in the output. Third, a multidimensional spatial occupancy grid map is created within the prediction time domain, and the trajectories within the prediction time domain are projected into the MSOGM. The uncertainty of the predicted trajectories is also represented using a Gaussian distribution and projected into the MSOGM. Fourth, a conflict check system is established for predicted trajectory pairs. Conflict checks are performed on trajectory pairs using the lateral and longitudinal safety distances calculated using the RSS model, and information related to conflicting trajectory pairs is screened and recorded. Fifth, scenario analysis is performed on the vehicles associated with conflicting trajectory pairs, determining the subsequent driving behavior of the conflicting vehicles within the prediction time domain. The vehicle trajectories are then re-predicted based on the driving behavior and a pre-trained model. Sixth, a path that satisfies the constraints is planned within the MSOGM, which has undergone conflict check and re-prediction. The motion planning method for autonomous vehicles provided by the present invention effectively improves trajectory prediction accuracy, reliability, and credibility.

[0063] In a second aspect, an embodiment of the present application further provides a motion planning device for an autonomous driving vehicle based on trajectory re-prediction, comprising:

[0064] A preprocessing module is configured to determine preset features and preprocess the preset features;

[0065] A neural network module is configured to determine a deep neural network, wherein the deep neural network is used to predict the future time domain Δt futThe movement trajectories of other vehicles in the

[0066] A grid diagram module is configured to establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within;

[0067] a conflict checking module configured to check the trajectory pairs for conflicting behaviors, and to filter and record relevant information of the trajectory pairs with conflicting behaviors;

[0068] a re-prediction module configured to re-predict the screened trajectory pairs with conflicting behaviors;

[0069] The planning module is configured to plan a vehicle motion path according to the re-prediction result.

[0070] In a third aspect, an embodiment of the present application further provides a motion planning device for an autonomous driving vehicle based on trajectory re-prediction, comprising: a memory, a processor, and a user interface;

[0071] The memory is used to store computer programs;

[0072] The user interface is used to interact with the user;

[0073] The processor is used to read the computer program in the memory, and when the processor executes the computer program, it implements the automatic driving vehicle motion planning method provided by the present invention.

[0074] In a fourth aspect, an embodiment of the present application further provides a processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, and when the processor executes the computer program, the autonomous driving vehicle motion planning method provided by the present invention is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0076] Figure 1 This is one of the schematic diagrams of motion planning for an autonomous driving vehicle based on trajectory re-prediction provided in an embodiment of the present application;

[0077] Figure 2 The second schematic diagram of motion planning for an autonomous driving vehicle based on trajectory re-prediction provided in an embodiment of the present application;

[0078] Figure 3 Driving behavior modeling and driving direction diagram provided in the embodiment of this application;

[0079] Figure 4 A schematic diagram of a neural network for predicting other vehicle trajectories based on driving behavior provided in an embodiment of the present application;

[0080] Figure 5 A schematic diagram of the projection of the predicted trajectory in MSOGM provided in an embodiment of the present application;

[0081] Figure 6 This is a schematic diagram of the structure of a motion planning device for an autonomous driving vehicle based on trajectory re-prediction provided in an embodiment of the present application;

[0082] Figure 7 This is a second structural diagram of the autonomous driving vehicle motion planning device based on trajectory re-prediction provided in an embodiment of the present application. DETAILED DESCRIPTION

[0083] To make the objectives, technical solutions, and advantages of the present invention more apparent, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the embodiments described herein are merely some, rather than all, of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0084] The following are explanations of some of the words that appear in the text:

[0085] 1. In the embodiments of the present invention, the term "and / or" describes the association relationship between associated objects, indicating that three possible relationships exist. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0086] 2. In the embodiments of the present application, the term "plurality" refers to two or more than two, and other quantifiers are similar.

[0087] 3. DNN, Deep Neural Networks;

[0088] 4. MSOGM, Multidimensional Space Occupancy GridMapping;

[0089] 5. RSS, Responsibility Sensitive Safety model;

[0090] 6. Other vehicles, i.e., vehicles other than the vehicle implementing the embodiment of the present invention;

[0091] 7. In this application, re-forecasting and re-forecasting are the same concepts;

[0092] 8. In this application, the concept of "his car" and "other cars" is the same.

[0093] Planning methods based on instantaneous states refer to planning modules that consider only the current surrounding traffic environment and then seek continuous path points within a desired range that satisfy constraints. Examples include spatial search planning methods based on the Dijkstra and A* algorithms, the Rapidly Exploring Random Tree (RRT) algorithm, and the Hybrid A* algorithm. However, these planning methods fail to account for the dynamic changes of traffic participants, and the resulting planned paths may not meet safety requirements in future traffic scenarios.

[0094] Planning methods based on long-term predictions refer to planning modules that take into account changes in the surrounding traffic environment over a period of time in the future, such as the changing trends in the movement paths of traffic participants and the status of traffic lights. Paths with higher safety factors are selected by calculating collision probabilities or modeling in the spatiotemporal dimensions. Such methods can plan their own movement paths based on the evolution of traffic scenarios over a period of time in the future and are predictive. However, such methods have high requirements for prediction accuracy, and the accuracy of trajectory predictions for other traffic participants will directly affect the reliability and safety of the planned path.

[0095] To address the issues of low accuracy, low credibility, and difficulty measuring reliability in the aforementioned planning methods for predicting the trajectory of other vehicles, embodiments of the present invention propose a motion planning method for autonomous vehicles based on trajectory re-prediction. Based on a Responsibility Sensitive Safety (RSS) model, the reliability of the other vehicle trajectories output by the prediction system is checked, and trajectories whose reliability does not meet preset conditions are re-predicted. This improves the accuracy of trajectory prediction, as well as its reliability and credibility, ultimately enhancing the safety of the planned path for the autonomous vehicle.

[0096] The motion planning method for autonomous vehicles based on trajectory re-prediction provided by an embodiment of the present invention comprises six parts: the first part extracts relevant features from the perception module and performs pre-processing. The second part is to design a properly designed deep neural network that can output the driving behavior of surrounding vehicles, and the output includes corresponding confidence levels. The third part is to establish a multidimensional spatial occupancy grid map within the prediction time domain and project the trajectories within the prediction time domain into the MSOGM. The uncertainty of the predicted trajectories is also represented by a Gaussian distribution and projected into the MSOGM. The fourth part is to establish a conflict check system for predicted trajectory pairs. The conflict check is performed on trajectory pairs using the lateral and longitudinal safety distances calculated by the RSS model, and relevant information on trajectory pairs with conflicting behavior is screened and recorded. The fifth part is to perform scenario analysis on the vehicles associated with the conflicting trajectory pairs, namely, to determine the subsequent driving behavior of the vehicles with conflicting behavior within the prediction time domain and re-predict the vehicle trajectories based on the driving behavior and the pre-trained model. The sixth part is to plan a path that meets the constraints in the MSOGM after the conflict check and re-prediction.

[0097] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0098] It should be noted that the order in which the embodiments of the present application are presented only represents the order of precedence of the embodiments, and does not represent the advantages or disadvantages of the technical solutions provided by the embodiments.

[0099] Example 1

[0100] See also Figure 1 , a schematic diagram of a motion planning method for an autonomous driving vehicle based on trajectory re-prediction is provided in an embodiment of the present application, such as Figure 1 As shown, the method includes steps S101 to S106:

[0101] S101, determining preset features and preprocessing the preset features;

[0102] As a preferred example, the preset features include one or a combination of the following:

[0103] Longitudinal position x, lateral position y, velocity v, acceleration a, yaw angle yaw, distance to the vehicle in front d f 、Previous vehicle type T f , road type R. That is, the preset features include one or more of the above features.

[0104] As a preferred example, the preset features include all the above features and satisfy the following relationship:

[0105]

[0106] Among them, n represents the vehicle number, N represents the total number of surrounding vehicles, represents the longitudinal position, lateral position, speed, acceleration, yaw angle, distance to the preceding vehicle, preceding vehicle type and road type of vehicle numbered n at time t, respectively. past Represents the set historical time, past n represents Δt past Vectorized representation of the driving information of vehicle numbered n within a time period, where t0 represents the current moment and past represents the sum of historical driving information of all surrounding vehicles.

[0107] As a preferred example, the pre-processing includes: f , the road type R is numerically represented and the sum of historical driving information of all surrounding vehicles past is normalized. As another preferred example, the preprocessing may also include standardizing and / or normalizing the above-mentioned preset features to eliminate the influence of features of different dimensions.

[0108] S102: Determine a deep neural network, wherein the deep neural network is used to predict the future time domain Δt fut The movement trajectories of other vehicles in the

[0109] As a preferred example, the future time domain motion trajectory of vehicle numbered n is pre n for:

[0110]

[0111] Among them, X n ,Y n The vehicle numbered n is in the future time domain Δt fut Vector representation of vertical and horizontal positions, Action m Represents the driving behavior of other cars, M represents the number of preset discrete driving behaviors, m is the number of preset discrete driving behaviors, Model() represents different Action m The corresponding prediction model.

[0112] As a preferred example, the deep neural network provided by the embodiment of the present invention is as follows Figure 4As shown in the figure, Multilayer Perceptron (MLP) refers to the superposition of multi-layer neural networks (NN), and LSTM-Encoder and LSTM-Decoder represent the encoder (Encoder) and decoder (Decoder) neural network structures composed of basic LSTM units.

[0113] As a preferred example, the deep neural network provided by this embodiment can also be divided into the following sub-parts:

[0114] (21) Encoding module for historical information of surrounding vehicles;

[0115] (22) A prediction module based on the Long Short-Term Memory (LSTM) neural network, which pre-trains different models based on historical data of different driving behaviors;

[0116] (23) Recognition module for the behavior intention of surrounding vehicles;

[0117] (24) Combining the intentions and historical trajectories of surrounding vehicles, the trajectory prediction results under different driving behaviors (intentions) are output and assigned confidence levels.

[0118] S103: Establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within;

[0119] As a preferred example, establishing the multidimensional space occupancy grid network diagram in the future time domain includes: mapping the motion trajectories of the other vehicles into the multidimensional space occupancy grid diagram MSOGM; and performing Gaussian noise processing on the mapped trajectory points.

[0120] As a preferred example, the future time domain Δt fut The multi-dimensional space occupancy grid network diagram can also be composed of the following two parts:

[0121] (31) Establishing a multidimensional space within the prediction time domain;

[0122] (32) For the predicted trajectory, Gaussian noise is combined to represent the uncertainty of the prediction result.

[0123] As a preferred example, the future time domain Δt is established fut The multi-dimensional space occupancy grid network diagram is the motion trajectory of other vehicles obtained in S102 mapped into the MSOGM, such as Figure 5As shown in the figure, the two lines represent the trajectories of two vehicles in the prediction time domain, pre1 and pre2 represent the trajectory points of vehicles numbered 1 and 2 in the prediction time domain respectively. Then, Gaussian noise processing is performed on the mapped trajectory points to describe the uncertainty of the prediction results.

[0124] S104, performing conflict behavior checks on the trajectory pairs, screening and recording relevant information of the trajectory pairs with conflicting behaviors;

[0125] As a preferred example, this step can be divided into the following parts:

[0126] (41) Establishment of a safety distance model based on RSS. Using the RSS model, inputting model parameters based on specific circumstances, and calculating the horizontal and vertical safety distances that should be maintained;

[0127] (42) All track pairs are checked for collision, with the check criteria being the horizontal and vertical safety distances;

[0128] (43) Filter out trajectory pairs with conflicting behaviors and record the corresponding vehicle information.

[0129] As a preferred example, this step S104 may include:

[0130] Calculate the lateral and longitudinal safety distances that two vehicles need to maintain;

[0131] Among them, the rear car C r Should be with the front car C f Maintain longitudinal safety distance d 0,lon for:

[0132]

[0133] Among them, v r Represents the longitudinal velocity of the following vehicle at the initial moment, v f represents the longitudinal velocity of the preceding vehicle at the initial moment, ρ represents the reaction time, and a max,accel,lon Represents the maximum longitudinal forward acceleration of the vehicle, a min,brake,lon Represents the minimum longitudinal braking acceleration of the vehicle, a max,brake,lon Represents the maximum longitudinal braking acceleration of the vehicle;

[0134] Left car C l Should be with the right car C r Minimum lateral safety distance d 0,lat for:

[0135]

[0136] Among them, μ represents the minimum lateral distance between the two vehicles to maintain safety, v1 represents the initial lateral speed of the left vehicle, v2 represents the initial lateral speed of the right vehicle, and v 1,ρ represents the lateral velocity of the left vehicle after time ρ and v 1,ρ =v1+ρa max,accel,lat , a max,accel,lat Represents the maximum lateral acceleration of the vehicle, v 2,ρ represents the lateral velocity of the right vehicle after time ρ and v 2,ρ =v2-ρa max,accle,lat , a min,brake,lat Represents the minimum lateral acceleration of the vehicle.

[0137] As a preferred example, the conflict behavior check includes:

[0138] Conflict i,j The judgment basis is:

[0139]

[0140] When Conflict i,j A value of 1 indicates that the predicted trajectory pair has conflicting behavior;

[0141] in, They represent the longitudinal and lateral coordinates of the vehicle numbered i at time t, They represent the longitudinal and lateral coordinates of the vehicle numbered j at time t respectively; k1, k2 represent the safety distance coefficients and k1, k2∈R, where R represents a real number.

[0142] S105, re-predicting the selected trajectory pairs with conflicting behaviors;

[0143] In this embodiment, re-predicting the selected trajectory pairs with conflicting behaviors may include the following steps:

[0144] (51) Conduct conflict behavior checks on related vehicles with conflicting behaviors based on the basic principles of RSS, including driving principles such as never hitting other vehicles from behind and not cutting in recklessly;

[0145] (52) Output the driving behavior that the corresponding vehicle should take and assign a confidence level;

[0146] (53) Re-predict the original trajectory based on the corresponding driving behavior;

[0147] (54) The trajectory is checked for conflicting behaviors in a loop until there is no conflict or the predicted system time limit is exceeded.

[0148] As a preferred example, the re-predicting of the screened trajectory pairs with conflicting behaviors may also include:

[0149] For vehicles with conflicting behaviors, the corresponding driving behaviors are output in combination with the RSS principle:

[0150]

[0151] Action rea Represents the corrected driving behavior of a vehicle with conflicting behavior after RSS analysis. Represents the action corresponding to vehicle number n rea The trajectory of the next heavy forecast;

[0152] The re-predicted trajectory is re-projected into the multi-dimensional space occupancy grid map MSOGM to obtain a new MSOGM.

[0153] As another preferred example, after this step, the re-prediction time may be limited, such as Figure 2 As shown:

[0154] S1052: Determine whether the calculation time of the re-prediction exceeds the preset threshold, that is, determine whether the re-prediction calculation has timed out. If so, return to S101, that is, restart the motion planning process from the beginning; if not, execute S106.

[0155] That is, if the re-prediction is completed within the predetermined time range, the vehicle movement path is planned according to the re-prediction result; if the re-prediction is not completed within the predetermined time range, the re-prediction attempt is abandoned and the process returns to S101 to restart the movement planning process.

[0156] S106: Plan the vehicle's motion path based on the re-prediction result.

[0157] As a preferred example, this step can be divided into the following parts:

[0158] (61) Determine the restrictions of the planned route according to different driving purposes;

[0159] (62) Select the safest path from all planned paths that meet the requirements.

[0160] As a preferred example, in this step, planning the vehicle motion path according to the re-prediction result may include:

[0161] Planning a vehicle motion path according to the new MSOGM, a preset objective function and constraints;

[0162] Wherein, the new MSOGM is:

[0163] MSOGM=[X ts ,Y ts ,T ts ]

[0164] X ts =[x k ],k∈[x min ,x max ]

[0165] Y ts =[y l ],l∈[y miin ,y max ]

[0166] T ts =[t o ],o∈[t0,t0+Δt fut ]

[0167] Among them, X ts ,Y ts ,T ts The vector representation of the vertical coordinate dimension, horizontal coordinate dimension and prediction time domain dimension of the MSOGM is represented in sequence, x k is the vertical position, y l is the horizontal position, t o is the coordinate scale on the time dimension, x min ,x max Represents the minimum, maximum, and y dimensions respectively. min ,y max Represent the minimum and maximum values of the y dimension, t0, t0+Δt respectively pre Represent the minimum and maximum values of the t dimension respectively;

[0168] The preset objective function and constraints are:

[0169]

[0170] path risk Represents the comprehensive risk value of each path, represents the longitudinal distance between the vehicle and the nearest neighbor vehicle j at time t, represents the lateral distance between the vehicle and the nearest neighbor vehicle j at time t, Respectively represent the longitudinal position and lateral position of the vehicle at time t, They represent the longitudinal position and lateral position of the nearest neighbor vehicle j at time t, respectively. k and q represent the distance risk coefficients, and k, q∈R, where R is a real number.

[0171] That is, this step finds a solution that satisfies the objective function and the constraints in the MSOGM, which is the replanned motion path.

[0172] Through the method of this embodiment, the input features required by the DNN are first obtained, and then the feature data is preprocessed, and a reasonable DNN is designed to output corresponding trajectories for different driving behaviors of other vehicles, and each trajectory is given a corresponding confidence level to compensate for the uncertainty of the prediction itself. Secondly, an MSOGM is established and the vehicle posture and the trajectory in the predicted time domain are projected into the MSOGM. At the same time, the predicted trajectory is refitted with the uncertainty of the trajectory point distribution using a Gaussian distribution and projected into the MSOGM together. Then, all trajectory pairs are checked for conflicting behaviors based on the RSS to filter out predicted trajectories that do not conform to the real scene, and the conflicting trajectory pairs are re-predicted based on the real scene to improve the reliability and objectivity of the prediction results. Finally, in the MSOGM, a path that meets the conditions is found according to the constraints of the plan, thereby improving the accuracy of the trajectory prediction, and improving reliability and credibility.

[0173] Example 2

[0174] Based on the same inventive concept, an embodiment of the present invention further provides a motion planning device for an autonomous driving vehicle based on trajectory re-prediction, such as Figure 6 As shown, the device includes:

[0175] The preprocessing module 601 is configured to determine preset features and preprocess the preset features;

[0176] The neural network module 602 is configured to determine a deep neural network for predicting the future time domain Δt fut The movement trajectories of other vehicles in the

[0177] The grid diagram module 603 is configured to establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within;

[0178] The conflict checking module 604 is configured to check the trajectory pairs for conflicting behaviors, and to filter and record relevant information of the trajectory pairs with conflicting behaviors;

[0179] a re-prediction module 605 , configured to re-predict the screened trajectory pairs with conflicting behaviors;

[0180] The planning module 606 is configured to plan the vehicle movement path according to the re-prediction result.

[0181] As a preferred example, the preprocessing module 601 is further configured to preprocess the preset features in the following manner:

[0182] The preset features include one or a combination of the following:

[0183] Longitudinal position x, lateral position y, velocity v, acceleration a, yaw angle yaw, distance to the vehicle in front d f 、Previous vehicle type T f , road type R;

[0184] The preset characteristics satisfy the following relationship:

[0185]

[0186] Among them, n represents the vehicle number, N represents the total number of surrounding vehicles, represents the longitudinal position, lateral position, speed, acceleration, yaw angle, distance to the preceding vehicle, preceding vehicle type and road type of vehicle numbered n at time t, respectively. past Represents the set historical time, past n represents Δt past Vectorized representation of the driving information of vehicle numbered n within a time period, where t0 represents the current moment and past represents the sum of historical driving information of all surrounding vehicles.

[0187] The pretreatment includes:

[0188] The preceding vehicle type T f , the road type R is numerically represented and the sum of historical driving information past of all surrounding vehicles is standardized.

[0189] As a preferred example, the neural network module 602 is further configured to predict the future time domain Δt according to the following method: fut The movement trajectories of other vehicles within:

[0190] The future motion trajectory of vehicle numbered n is pre n for:

[0191] pre n =[X n ,Y n ]=Model(past n ,Action m )

[0192]

[0193] Action m =[1,2,...,M],M≥1

[0194] Among them, X n ,Y n The vehicle numbered n is in the future time domain Δt fut Vector representation of vertical and horizontal positions, Action mRepresents the driving behavior of other cars, M represents the number of preset discrete driving behaviors, m is the number of preset discrete driving behaviors, Model() represents different Action m The corresponding prediction model.

[0195] As a preferred example, the grid map module 603 is further configured to establish a multi-dimensional space occupancy grid network map in the future time domain according to the following method:

[0196] Mapping the motion trajectories of the other vehicles into a multi-dimensional space occupancy grid map MSOGM;

[0197] Perform Gaussian noise processing on the mapped trajectory points.

[0198] As a preferred example, the conflict checking module 604 is further configured to perform conflict behavior checks on the trajectory pairs, and filter and record relevant information of the trajectory pairs with conflicting behaviors according to the following method:

[0199] Calculate the lateral and longitudinal safety distances that two vehicles need to maintain;

[0200] Among them, the rear car C r Should be with the front car C f Maintain longitudinal safety distance d 0,lon for:

[0201]

[0202] Among them, v r Represents the longitudinal velocity of the following vehicle at the initial moment, v f represents the longitudinal velocity of the preceding vehicle at the initial moment, ρ represents the reaction time, and a max,accel,lon Represents the maximum longitudinal forward acceleration of the vehicle, a min,brake,lon Represents the minimum longitudinal braking acceleration of the vehicle, a max,brake,lon Represents the maximum longitudinal braking acceleration of the vehicle;

[0203] Left car C l Should be with the right car C r Maintain the minimum lateral safety distance d 0,lat for:

[0204]

[0205] Among them, μ represents the minimum lateral distance between the two vehicles to maintain safety, v1 represents the initial lateral speed of the left vehicle, v2 represents the initial lateral speed of the right vehicle, and v 1,ρ represents the lateral velocity of the left vehicle after time ρ and v 1,ρ =v1+ρa max,accel,lat , a max,accel,lat Represents the maximum lateral acceleration of the vehicle, v2,ρ represents the lateral velocity of the right vehicle after time ρ and v 2,ρ =v2-ρa max,accle,lat , a min,brake,lat Represents the minimum lateral acceleration of the vehicle.

[0206] As a preferred example, the conflict checking module 604 is further configured to perform conflict behavior checking on the trajectory pair according to the following method:

[0207] Conflict i,j The judgment basis is:

[0208]

[0209] When Conflict i,j A value of 1 indicates that the predicted trajectory pair has conflicting behavior;

[0210] in, They represent the longitudinal and lateral coordinates of the vehicle numbered i at time t, They represent the longitudinal and lateral coordinates of the vehicle numbered j at time t respectively; k1, k2 represent the safety distance coefficients and k1, k2∈R, where R represents a real number.

[0211] As a preferred example, the re-prediction module 605 is further configured to re-predict the selected trajectory pairs with conflicting behaviors according to the following method:

[0212] For vehicles with conflicting behaviors, the corresponding driving behaviors are output in combination with the RSS principle:

[0213]

[0214] Action rea Represents the corrected driving behavior of a vehicle with conflicting behavior after RSS analysis. Represents the action corresponding to vehicle number n rea The trajectory of the next heavy forecast;

[0215] The re-predicted trajectory is re-projected into the multi-dimensional space occupancy grid map MSOGM to obtain a new MSOGM.

[0216] As a preferred example, the planning module 606 is further configured to plan the vehicle motion path according to the re-prediction result:

[0217] Planning a vehicle motion path according to the new MSOGM, a preset objective function and constraints;

[0218] Wherein, the new MSOGM is:

[0219] MSOGM=[X ts ,Y ts ,T ts ]

[0220] X ts =[x k ],k∈[x min ,x max ]

[0221] Y ts =[y l ],l∈[y miin ,y max ]

[0222] T ts =[t o ],o∈[t0,t0+Δt fut ]

[0223] Among them, X ts ,Y ts ,T ts The vector representation of the vertical coordinate dimension, horizontal coordinate dimension and prediction time domain dimension of the MSOGM is represented in sequence, x k is the vertical position, y l is the horizontal position, t o is the coordinate scale on the time dimension, x min ,x max Represents the minimum, maximum, and y dimensions respectively. min ,y max Represent the minimum and maximum values of the y dimension, t0, t0+Δt respectively pre Represent the minimum and maximum values of the t dimension respectively;

[0224] The preset objective function and constraints are:

[0225]

[0226] path risk Represents the comprehensive risk value of each path, represents the longitudinal distance between the vehicle and the nearest neighbor vehicle j at time t, represents the lateral distance between the vehicle and the nearest neighbor vehicle j at time t, Respectively represent the longitudinal position and lateral position of the vehicle at time t, They represent the longitudinal position and lateral position of the nearest neighbor vehicle j at time t, respectively. k and q represent the distance risk coefficients, and k, q∈R, where R is a real number.

[0227] It should be noted that the pre-processing module 601 provided in this embodiment can implement all the functions included in step S101 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0228] It should be noted that the neural network module 602 provided in this embodiment can implement all the functions included in step S102 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0229] It should be noted that the grid map module 603 provided in this embodiment can implement all the functions included in step S103 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0230] It should be noted that the conflict checking module 604 provided in this embodiment can implement all the functions included in step S104 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0231] It should be noted that the re-prediction module 605 provided in this embodiment can implement all the functions included in steps S105 and S1052 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0232] It should be noted that the planning module 606 provided in this embodiment can implement all the functions included in step S106 in the first embodiment, solve the same technical problems, and achieve the same technical effects, and will not be described in detail here.

[0233] It should be noted that the device provided in Example 2 and the method provided in Example 1 belong to the same inventive concept, solve the same technical problems, and achieve the same technical effects. The device provided in Example 2 can implement all the methods in Example 1, and the similarities will not be repeated here.

[0234] Example 3

[0235] Based on the same inventive concept, an embodiment of the present invention further provides a motion planning device for an autonomous driving vehicle based on trajectory re-prediction, such as Figure 7 As shown, the device includes:

[0236] It includes a memory 702, a processor 701 and a user interface 703;

[0237] The memory 702 is used to store computer programs;

[0238] The user interface 703 is used to interact with the user;

[0239] The processor 701 is configured to read the computer program in the memory 702. When the processor 701 executes the computer program, the following steps are implemented:

[0240] Determining preset features and preprocessing the preset features;

[0241] Determine a deep neural network for predicting the future time domain Δt fut The movement trajectories of other vehicles in the

[0242] Establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within;

[0243] Check the trajectory pairs for conflicting behaviors, filter and record relevant information of trajectory pairs with conflicting behaviors;

[0244] Re-predict the selected trajectory pairs with conflicting behaviors;

[0245] Plan the vehicle's motion path based on the re-prediction results.

[0246] Among them, Figure 7 In the present invention, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits such as one or more processors represented by processor 701 and memory represented by memory 702. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore not further described herein. The bus interface provides an interface. Processor 701 is responsible for managing the bus architecture and general processing, while memory 702 can store data used by processor 701 when performing operations.

[0247] The processor 701 may be a CPU, an ASIC, an FPGA or a CPLD, and the processor 501 may also adopt a multi-core architecture.

[0248] When the processor 701 executes the computer program stored in the memory 702, it implements any of the autonomous driving vehicle motion planning methods based on trajectory re-prediction in Example 1.

[0249] It should be noted that the device provided in Example 3 and the method provided in Example 1 belong to the same inventive concept, solve the same technical problems, and achieve the same technical effects. The device provided in Example 3 can implement all the methods in Example 1, and the similarities will not be repeated here.

[0250] The present application also provides a processor-readable storage medium, wherein the processor-readable storage medium stores a computer program, and when the processor executes the computer program, it implements any of the autonomous driving vehicle motion planning methods based on trajectory re-prediction in Example 1.

[0251] It should be noted that the division of units in the embodiments of the present application is schematic and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0252] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage and optical storage, etc.) that contain computer-usable program code.

[0253] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0254] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0255] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A motion planning method for an autonomous driving vehicle based on trajectory re-prediction, characterized in that: include: Determining preset features and preprocessing the preset features; Determine a deep neural network for predicting the future time domain Δt fut The movement trajectories of other vehicles in the Establish the future time domain △t fut Multi-dimensional space occupancy grid network diagram within; Check the trajectory pairs for conflicting behaviors, filter and record relevant information of trajectory pairs with conflicting behaviors; Re-predict the selected trajectory pairs with conflicting behaviors; Plan the vehicle's motion path based on the re-prediction results; The step of checking for conflicting behaviors on the trajectory pairs and screening and recording relevant information of the trajectory pairs with conflicting behaviors includes: Calculate the lateral and longitudinal safety distances that two vehicles need to maintain; Among them, the rear car C r Should be with the front car C f Maintain longitudinal safety distance d 0,lon for: Among them, v r Represents the longitudinal velocity of the following vehicle at the initial moment, v f represents the longitudinal velocity of the preceding vehicle at the initial moment, ρ represents the reaction time, and a max,accel,lon Represents the maximum longitudinal forward acceleration of the vehicle, a min,brake,lon Represents the minimum longitudinal braking acceleration of the vehicle, a max,brake,lon Represents the maximum longitudinal braking acceleration of the vehicle; Left car C l Should be with the right car C r Maintain the minimum lateral safety distance d 0,lat for: Among them, μ represents the minimum lateral distance between the two vehicles to maintain safety, v1 represents the initial lateral speed of the left vehicle, v2 represents the initial lateral speed of the right vehicle, and v 1,ρ represents the lateral velocity of the left vehicle after time ρ and v 1,ρ =v1+ρa max,accel,lat , a max,accel,lat Represents the maximum lateral acceleration of the vehicle, v 2,ρ represents the lateral velocity of the right vehicle after time ρ and v 2,ρ =v2-ρa max,accle,lat , a min,brake,lat Represents the minimum lateral acceleration of the vehicle.

2. The method according to claim 1, characterized in that The determining of the preset features and preprocessing the preset features includes: The preset features include one or a combination of the following: Longitudinal position x, lateral position y, velocity v, acceleration a, yaw angle yaw, distance to the vehicle in front d f 、Previous vehicle type T f , road type R; The pretreatment includes: The preceding vehicle type T f , the road type R is numerically represented and the sum of historical driving information past of all surrounding vehicles is standardized.

3. The method according to claim 2, characterized in that The preset characteristics satisfy the following relationship: Among them, n represents the vehicle number, N represents the total number of surrounding vehicles, They represent the longitudinal position, lateral position, speed, acceleration, yaw angle, distance to the preceding vehicle, preceding vehicle type and road type of vehicle numbered n at time t, respectively. past Represents the set historical time, past n Indicates △t past Vectorized representation of the driving information of vehicle numbered n within a time period, where t0 represents the current moment and past represents the sum of historical driving information of all surrounding vehicles.

4. The method according to claim 3, characterized in that The deep neural network is used to predict the future time domain Δt fut The motion trajectories of other vehicles in the include: The future motion trajectory of vehicle numbered n is pre n for: Among them, X n ,Y n The vehicle numbered n is in the future time domain △t fut Vector representation of vertical and horizontal positions, Action m Represents the driving behavior of other cars, M represents the number of preset discrete driving behaviors, m is the number of preset discrete driving behaviors, Model() represents different Action m The corresponding prediction model.

5. The method according to claim 4, characterized in that The establishment of the future time domain Δt fut The multi-dimensional space occupancy grid network diagram includes: Mapping the motion trajectories of the other vehicles into a multi-dimensional space occupancy grid map MSOGM; Perform Gaussian noise processing on the mapped trajectory points.

6. The method according to claim 1, characterized in that The conflict behavior check on the trajectory pair includes: Conflict i,j The judgment basis is: t∈[t0,t0+△t fut ] When Conflict i,j A value of 1 indicates that the predicted trajectory pair has conflicting behavior; in, They represent the longitudinal and lateral coordinates of the vehicle numbered i at time t, They represent the longitudinal and lateral coordinates of the vehicle numbered j at time t respectively; k1, k2 represent the safety distance coefficients and k1, k2∈R, where R represents a real number.

7. The method according to claim 6, characterized in that The re-predicting of the selected trajectory pairs with conflicting behaviors includes: For vehicles with conflicting behaviors, the corresponding driving behaviors are output in combination with the RSS principle: Action rea Represents the corrected driving behavior of a vehicle with conflicting behavior after RSS analysis. Represents the action corresponding to vehicle number n rea The trajectory of the next heavy forecast; The re-predicted trajectory is re-projected into the multi-dimensional space occupancy grid map MSOGM to obtain a new MSOGM.

8. The method according to claim 7, characterized in that Also includes: If the calculation time of the re-prediction exceeds the preset threshold, the re-prediction is terminated.

9. The method according to claim 7, characterized in that Planning the vehicle motion path according to the re-prediction result includes: Planning a vehicle motion path according to the new MSOGM, a preset objective function and constraints; Wherein, the new MSOGM is: MSOGM=[X ts ,Y ts ,T ts ] X ts =[x k ],k∈[x min ,x max ] AND ts =[and l ],l∈[and miin ,and max ] T ts =[t o ],o∈[t0,t0+△t fut ] Among them, X ts ,Y ts ,T ts The vector representation of the vertical coordinate dimension, horizontal coordinate dimension and prediction time domain dimension of the MSOGM is represented in sequence, x k is the vertical position, y l is the horizontal position, t o is the coordinate scale on the time dimension, x min ,x max Represents the minimum, maximum, and y dimensions respectively. min ,y max Represent the minimum and maximum values of the y dimension, t0, t0+△t pre Represent the minimum and maximum values of the t dimension respectively; The preset objective function and constraints are: path risk Represents the comprehensive risk value of each path, represents the longitudinal distance between the vehicle and the nearest neighbor vehicle j at time t, represents the lateral distance between the vehicle and the nearest neighbor vehicle j at time t, Respectively represent the longitudinal position and lateral position of the vehicle at time t, They represent the longitudinal position and lateral position of the nearest neighbor vehicle j at time t, respectively. k and q represent the distance risk coefficients, and k, q∈R, where R is a real number.

10. A motion planning device for an autonomous driving vehicle based on trajectory re-prediction, characterized in that: include: A preprocessing module is configured to determine preset features and preprocess the preset features; A neural network module is configured to determine a deep neural network for predicting the future time domain Δt fut The movement trajectories of other vehicles in the A grid diagram module is configured to establish the future time domain Δt fut Multi-dimensional space occupancy grid network diagram within; a conflict checking module configured to check the trajectory pairs for conflicting behaviors, and to filter and record relevant information of the trajectory pairs with conflicting behaviors; a re-prediction module configured to re-predict the screened trajectory pairs with conflicting behaviors; a planning module configured to plan a vehicle motion path according to a result of the re-prediction; The step of checking for conflicting behaviors on the trajectory pairs and screening and recording relevant information of the trajectory pairs with conflicting behaviors includes: Calculate the lateral and longitudinal safety distances that two vehicles need to maintain; Among them, the rear car C r Should be with the front car C f Maintain longitudinal safety distance d 0,lon for: Among them, v r Represents the longitudinal velocity of the following vehicle at the initial moment, v f represents the longitudinal velocity of the preceding vehicle at the initial moment, ρ represents the reaction time, and a max,accel,lon Represents the maximum longitudinal forward acceleration of the vehicle, a min,brake,lon Represents the minimum longitudinal braking acceleration of the vehicle, a max,brake,lon Represents the maximum longitudinal braking acceleration of the vehicle; Left car C l Should be with the right car C r Maintain the minimum lateral safety distance d 0,lat for: Among them, μ represents the minimum lateral distance between the two vehicles to maintain safety, v1 represents the initial lateral speed of the left vehicle, v2 represents the initial lateral speed of the right vehicle, and v 1,ρ represents the lateral velocity of the left vehicle after time ρ and v 1,ρ =v1+ρa max,accel,lat , a max,accel,lat Represents the maximum lateral acceleration of the vehicle, v 2,ρ represents the lateral velocity of the right vehicle after time ρ and v 2,ρ =v2-ρa max,accle,lat , a min,brake,lat Represents the minimum lateral acceleration of the vehicle.

11. A motion planning device for an autonomous driving vehicle based on trajectory re-prediction, characterized in that: including a memory, a processor, and a user interface; The memory is used to store computer programs; The user interface is used to interact with the user; The processor is used to read the computer program in the memory, and when the processor executes the computer program, it implements the autonomous driving vehicle motion planning method based on trajectory re-prediction as described in one of claims 1 to 9.

12. A processor-readable storage medium, characterized in that: The processor-readable storage medium stores a computer program, and when the processor executes the computer program, the autonomous driving vehicle motion planning method based on trajectory re-prediction as described in one of claims 1 to 9 is implemented.

Images