Vehicle trajectory prediction method and device, server and storage medium

Abstract

The invention relates to a vehicle trajectory prediction method and device, a server and a storage medium, and the method comprises the steps: receiving trajectory points uploaded by a plurality of vehicles, and obtaining a historical fitting trajectory of each vehicle based on the trajectory points; obtaining a transverse prediction trajectory and a longitudinal prediction trajectory of each vehicle based on the historical fitting trajectory and the variation of the trajectory points; screening out at least two collision risk vehicles with collision risks based on the transverse prediction trajectory and the longitudinal prediction trajectory of each vehicle; the state matrix stability of each collision risk vehicle is calculated in sequence, and the transverse prediction track and the longitudinal prediction track of each collision risk vehicle are corrected according to the state matrix stability. Therefore, the technical problems that in the prior art, adaptability to environment changes is poor, dependence on high-quality driving data is achieved, and sudden road conditions are difficult to deal with are solved.

Description

Technical Field

[0001] The present application relates to the technical field of trajectory prediction, and in particular to a vehicle trajectory prediction method, device, server and storage medium. Background Art

[0002] Vehicle trajectory prediction is an important component in Intelligent Transportation Systems (ITS), which involves predicting the future path or position of a vehicle. This technology is crucial for autonomous vehicles, traffic flow management, accident prevention, etc.

[0003] In the related technology, the method based on a single physical model constructs a model through physical parameters such as acceleration, speed and vehicle dynamic characteristics, and obtains a trajectory prediction result with high robustness, but has poor adaptability to environmental changes and is not conducive to practical application; the method based on behavioral cloning can predict the vehicle trajectory by imitating the behavior of human drivers, so that the vehicle's automatic control behavior can be more in line with human logic in subsequent trajectory applications, but it relies on high-quality human driving data and may not be able to handle rare events or emergencies well.

[0004] In summary, the relevant technologies have poor adaptability to environmental changes, rely on high-quality driving data, and have difficulty coping with sudden road conditions, which needs to be improved. Summary of the invention

[0005] The present application provides a vehicle trajectory prediction method, device, vehicle and storage medium to solve the technical problems in related technologies that are poorly adaptable to environmental changes, rely on high-quality driving data, and are difficult to cope with sudden road conditions.

[0006] A first aspect of the present application provides a vehicle trajectory prediction method, which is applied to a server, wherein the method includes the following steps: receiving trajectory points uploaded by multiple vehicles, and obtaining a historical fitting trajectory of each vehicle based on the trajectory points; obtaining a lateral predicted trajectory and a longitudinal predicted trajectory of each vehicle based on the historical fitting trajectory and a change in the trajectory points; screening out at least two collision risk vehicles with a collision risk based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle; calculating the state matrix stability of each collision risk vehicle in turn, and correcting the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the state matrix stability.

[0007] Optionally, in one embodiment of the present application, the lateral predicted trajectory of each vehicle is obtained based on the historical fitting trajectory and the change in the trajectory points, including: determining a corresponding target reference line according to the vehicle center, current driving lane and lane width of each vehicle; obtaining a first state matrix based on a straight-ahead model and a turning model for each vehicle based on the target reference line, the vehicle center and a preset transfer probability; and obtaining the lateral predicted trajectory of the vehicle based on the first state matrix and the historical fitting trajectory of the vehicle.

[0008] Optionally, in one embodiment of the present application, determining the corresponding target reference line based on the vehicle center, the current driving lane and the lane width of each vehicle includes: calculating the lateral offset between the vehicle center and the lane centerline of the current driving lane; when the lateral offset is less than a preset offset threshold, determining that the target reference line is the lane centerline; when the lateral offset is greater than or equal to the preset offset threshold, determining whether the lateral offset is greater than a preset width threshold, if the lateral offset is greater than the preset width threshold, determining that the target reference line is the lane centerline of the adjacent lane, wherein the preset width threshold is obtained from the lane width.

[0009] Optionally, in one embodiment of the present application, after obtaining the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle based on the historical fitting trajectory and the change in the trajectory points, it also includes: screening out all lane-departing vehicles that meet a preset cross-border risk based on the lateral predicted trajectory and the current driving lane; and correcting the lateral predicted trajectory of the lane-departing vehicle based on the lane boundary of the current driving lane.

[0010] Optionally, in one embodiment of the present application, the method of obtaining the predicted longitudinal trajectory of each vehicle based on the historical fitting trajectory and the change in the trajectory points includes: obtaining the displacement of each vehicle in each preset sampling period and obtaining the corresponding displacement change rate; obtaining a second state matrix of the vehicle based on a uniform speed model and a uniform acceleration model based on the displacement change rate and a preset transition probability; and obtaining the predicted longitudinal trajectory of the vehicle based on the second state matrix and the historical fitting trajectory.

[0011] Optionally, in one embodiment of the present application, the correction of the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the state matrix stability includes: setting all collision risk vehicles whose lateral predicted trajectory and longitudinal predicted trajectory intersect at the same position point as a risk collision group; determining the collision risk vehicle with the largest state matrix stability in the risk collision group as an avoidance vehicle, and correcting the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle; judging whether there is a collision risk between the remaining collision risk vehicles based on the lateral predicted trajectory and the longitudinal predicted trajectory of the remaining collision risk vehicles in the risk collision group; if there is the collision risk between the remaining collision risk vehicles, determining the remaining collision risk vehicle with the largest state matrix stability among the remaining collision risk vehicles as the avoidance vehicle, and correcting the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle until there is no collision risk between the remaining collision risk vehicles.

[0012] Optionally, in one embodiment of the present application, after correcting the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle, it also includes: sending the lateral predicted trajectory and the longitudinal predicted trajectory to a corresponding vehicle so that the corresponding vehicle performs corresponding automatic driving control based on the lateral predicted trajectory and the longitudinal predicted trajectory; sending the corrected lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding collision risk vehicle so that the corresponding collision risk vehicle performs corresponding automatic driving control based on the corrected lateral predicted trajectory and the longitudinal predicted trajectory.

[0013] A second aspect of the present application provides a vehicle trajectory prediction device, which is applied to a server, wherein the device includes: a receiving module, which is used to receive trajectory points uploaded by multiple vehicles, and obtain a historical fitting trajectory of each vehicle based on the trajectory points; a prediction module, which is used to obtain a lateral predicted trajectory and a longitudinal predicted trajectory of each vehicle based on the historical fitting trajectory and the change in the trajectory points; a first screening module, which is used to screen out at least two collision risk vehicles with a collision risk based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle; and a first correction module, which is used to calculate the state matrix stability of each collision risk vehicle in turn, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the state matrix stability.

[0014] Optionally, in one embodiment of the present application, the prediction module includes: a determination unit, used to determine the corresponding target reference line according to the vehicle center, the current driving lane and the lane width of each vehicle; a first calculation unit, used to obtain the first state matrix based on the straight driving model and the turning model of each vehicle based on the target reference line, the vehicle center and the preset transfer probability; a first prediction unit, used to obtain the lateral predicted trajectory of the vehicle based on the first state matrix and the historical fitting trajectory of the vehicle.

[0015] Optionally, in one embodiment of the present application, the determination unit includes: a calculation subunit, used to calculate the lateral offset between the center of the vehicle and the lane centerline of the current driving lane; a first determination subunit, used to determine that the target reference line is the lane centerline when the lateral offset is less than a preset offset threshold; a second determination subunit, used to determine whether the lateral offset is greater than a preset width threshold when the lateral offset is greater than or equal to the preset offset threshold, and if the lateral offset is greater than the preset width threshold, determine that the target reference line is the lane centerline of the adjacent lane, wherein the preset width threshold is obtained from the lane width.

[0016] Optionally, in one embodiment of the present application, it also includes: a second screening module, used to screen out all lane-departing vehicles that meet a preset cross-border risk based on the lateral predicted trajectory and the current driving lane; and a second correction module, used to correct the lateral predicted trajectory of the lane-departing vehicle based on the lane boundary of the current driving lane.

[0017] Optionally, in one embodiment of the present application, the prediction module includes: an acquisition unit, used to acquire the displacement of each vehicle in each preset sampling period and obtain the corresponding displacement change rate; a second calculation unit, used to obtain the second state matrix of the vehicle based on the uniform speed model and the uniform acceleration model based on the displacement change rate and the preset transfer probability; a second prediction unit, used to obtain the longitudinal predicted trajectory of the vehicle based on the second state matrix and the historical fitting trajectory.

[0018] Optionally, in one embodiment of the present application, the correction module includes: a setting unit, used to set all collision risk vehicles whose lateral predicted trajectories and longitudinal predicted trajectories intersect at the same position point as a risk collision group; a first correction unit, used to determine that the collision risk vehicle with the largest state matrix stability in the risk collision group is an avoidance vehicle, and correct the lateral predicted trajectory and longitudinal predicted trajectory of the avoidance vehicle; a judgment unit, used to judge whether there is a collision risk between the remaining collision risk vehicles based on the lateral predicted trajectories and longitudinal predicted trajectories of the remaining collision risk vehicles in the risk collision group; a second correction unit, used to determine that the remaining collision risk vehicle with the largest state matrix stability among the remaining collision risk vehicles is an avoidance vehicle when the collision risk exists between the remaining collision risk vehicles, and correct the lateral predicted trajectory and longitudinal predicted trajectory of the avoidance vehicle until there is no collision risk between the remaining collision risk vehicles.

[0019] Optionally, in one embodiment of the present application, it also includes: a first generating module, used to send the lateral predicted trajectory and the longitudinal predicted trajectory to a corresponding vehicle, so that the corresponding vehicle performs corresponding automatic driving control based on the lateral predicted trajectory and the longitudinal predicted trajectory; a second sending module, used to send the corrected lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding collision risk vehicle, so that the corresponding collision risk vehicle performs corresponding automatic driving control based on the corrected lateral predicted trajectory and the longitudinal predicted trajectory.

[0020] The third aspect of the present application provides a server, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the vehicle trajectory prediction method as described in the above embodiment.

[0021] A fourth aspect of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the vehicle trajectory prediction method as described in the above embodiment.

[0022] The fifth aspect of the present application provides a computer program product, including a computer program, which, when executed, is used to implement the above vehicle trajectory prediction method.

[0023] The embodiment of the present application can obtain the historical fitting trajectory of each vehicle based on the trajectory points uploaded by multiple vehicles, and obtain the lateral predicted trajectory and longitudinal predicted trajectory of each vehicle based on the historical fitting trajectory and the change in the trajectory points, and then screen out at least two collision risk vehicles with collision risk, and correct the lateral predicted trajectory and longitudinal predicted trajectory of each collision risk vehicle according to the stability of the state matrix of each collision risk vehicle, so as to achieve vehicle trajectory prediction, and fully consider the target vehicle as an inertial body, whose motion law conforms to the basic mechanical constraints. The result of trajectory prediction is fused and calculated according to the interactive multi-model algorithm, and based on the detection of collision possibility, it is evaluated according to the state stability of the vehicle to determine which one is more inclined to avoid collision, and it can cope with environmental changes and sudden road conditions, so as to serve as a reference for the automatic driving control of each vehicle. Thus, the technical problems of poor adaptability to environmental changes, dependence on high-quality driving data, and difficulty in coping with sudden road conditions in the related technology are solved.

[0024] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0026] Figure 1 A flowchart of a vehicle trajectory prediction method provided according to an embodiment of the present application;

[0027] Figure 2 A schematic diagram of the principle of lateral trajectory prediction provided according to an embodiment of the present application;

[0028] Figure 3 A schematic diagram of the principle of longitudinal trajectory prediction provided according to an embodiment of the present application;

[0029] Figure 4 A schematic diagram of the collision correction principle provided according to an embodiment of the present application;

[0030] Figure 5 A schematic diagram of the boundary correction principle provided according to an embodiment of the present application;

[0031] Figure 6 A schematic diagram of the structure of a vehicle trajectory prediction device provided according to an embodiment of the present application;

[0032] Figure 7 A schematic diagram of the structure of a server provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0034] The following describes the vehicle trajectory prediction method, device, vehicle and storage medium of the embodiment of the present application with reference to the accompanying drawings. In view of the poor adaptability to environmental changes in the related technologies mentioned in the above background technology, and the technical problems that it is difficult to cope with sudden road conditions, the present application provides a vehicle trajectory prediction method, in which the historical fitting trajectory of each vehicle can be obtained according to the trajectory points uploaded by multiple vehicles, and the lateral prediction trajectory and longitudinal prediction trajectory of each vehicle can be obtained based on the change in the historical fitting trajectory and the trajectory point, and then at least two collision risk vehicles with collision risk are screened out, and the lateral prediction trajectory and longitudinal prediction trajectory of each collision risk vehicle are corrected according to the state matrix stability of each collision risk vehicle, so as to realize vehicle trajectory prediction, and fully consider the target vehicle as an inertial body, and its motion law conforms to the basic mechanical constraints. The result of trajectory prediction is fused and calculated according to the interactive multi-model algorithm, and based on the detection of collision possibility, it is evaluated according to the state stability of the vehicle which is more inclined to avoid collision, and it can cope with environmental changes and sudden road conditions, so as to serve as a reference for the automatic driving control of each vehicle. Thus, the technical problems that the poor adaptability to environmental changes in the related technologies, the reliance on high-quality driving data, and the difficulty in coping with sudden road conditions are solved.

[0035] Specifically, Figure 1 A schematic flow chart of a vehicle trajectory prediction method provided in an embodiment of the present application.

[0036] like Figure 1 As shown, the vehicle trajectory prediction method is applied to a server, wherein the method comprises the following steps:

[0037] In step S101, trajectory points uploaded by multiple vehicles are received, and a history fitting trajectory of each vehicle is obtained based on the trajectory points.

[0038] In the actual implementation process, the embodiments of the present application can be applied to a server to achieve simultaneous monitoring of multiple vehicles and predict vehicle trajectories from a macro perspective, so that each vehicle only needs to detect its own data and upload it, reducing the computing burden of each vehicle.

[0039] The vehicle can upload its current location at regular sampling intervals. The server determines the corresponding trajectory points based on the current location uploaded by the vehicle, and obtains the corresponding historical fitting trajectory based on the trajectory points of each vehicle to describe the historical driving trajectory of each vehicle.

[0040] In step S102 , the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle are obtained based on the history matching trajectory and the change amount of the trajectory points.

[0041] Furthermore, the embodiments of the present application can perform lateral trajectory prediction and longitudinal trajectory prediction based on the historical fitting trajectory of each vehicle and the change amount of the trajectory points.

[0042] When performing lateral trajectory prediction, the change amount of the trajectory point may be the lateral offset, and when performing longitudinal trajectory prediction, the change rate of the trajectory point may be the longitudinal displacement change.

[0043] Optionally, in one embodiment of the present application, a lateral predicted trajectory of each vehicle is obtained based on the historical fitting trajectory and the change in trajectory points, including: determining a corresponding target reference line according to the vehicle center, current driving lane and lane width of each vehicle; obtaining a first state matrix based on a straight-line model and a turning model for each vehicle based on the target reference line, the vehicle center and a preset transfer probability; and obtaining a lateral predicted trajectory of the vehicle based on the first state matrix and the historical fitting trajectory of the vehicle.

[0044] The prediction of lateral trajectories is elaborated here.

[0045] The embodiment of the present application can use two models to predict the lateral direction of the vehicle, namely the straight-moving model S and the turning model T. In the straight-moving model, the vehicle travels along the center line of the current lane, and if it encounters a curve, it still keeps driving in the center of the lane. In the turning model, the vehicle turns to one side according to a constant turning radius.

[0046] First, the embodiment of the present application can determine the target reference line of the vehicle based on the vehicle center, current driving lane and lane width of each vehicle, that is, the target reference line is determined based on the numerical value between the offset distance between the vehicle and the lane centerline and the lane width.

[0047] After determining the target reference line, the embodiment of the present application can record the lateral offset between the center of the vehicle and the reference line as ΔL, continuously track the vehicle through the perception data of the autonomous driving vehicle, and monitor the value of ΔL in each sampling period;

[0048] If ΔL does not exceed the offset threshold for a period of time, the vehicle is considered to be in state S and the state matrix is ​​set, such as [S, T] = [0.9, 0.1];

[0049] If ΔL exceeds the offset threshold, the transition probabilities of the four gradients are set according to the magnitude of the threshold, such as P S-T ∈(1, 0.8, 0.6, 0.4). Update S and T values, S = S*P S-T, T = 1-S;

[0050] When the S value is lower than 0.5, the state matrix is ​​inverted [S, T] = [0.1, 0.9];

[0051] When ΔL<lane width (default 3.5m)*ζ, set the transition probability P of four gradients according to the deviation of ΔL T-S ∈(1, 0.8, 0.6, 0.4), update T and S values, T = T*P T-S , S=1-T.

[0052] If ΔL≥lane width (default 3.5m)*ε and lasts for a period of time, the target vehicle is considered to be in state S, and the state matrix is ​​reset to [S, T]=[0.1, 0.9], where ζ is the empirical coefficient.

[0053] When the T value is lower than 0.5, the state matrix is ​​inverted [S, T] = [0.9, 0.1].

[0054] The embodiment of the present application can fit the historical trajectory points of the vehicle into a curve and predict two trajectories for the vehicle, one is a straight trajectory 1 along the reference line of the target vehicle, and the other is a lane change trajectory 2 with a transverse curvature based on the historical trajectory curve of the vehicle. According to the state matrix [S, T], the final predicted trajectory is the result of the fusion of trajectory 1 and trajectory 2. The higher the value of S, the closer the final predicted trajectory is to trajectory 1, and vice versa.

[0055] Optionally, in one embodiment of the present application, a corresponding target reference line is determined based on the vehicle center, the current driving lane and the lane width of each vehicle, including: calculating the lateral offset between the vehicle center and the lane centerline of the current driving lane; when the lateral offset is less than a preset offset threshold, determining the target reference line to be the lane centerline; when the lateral offset is greater than or equal to the preset offset threshold, determining whether the lateral offset is greater than a preset width threshold, if the lateral offset is greater than the preset width threshold, determining the target reference line to be the lane centerline of the adjacent lane, wherein the preset width threshold is obtained from the lane width.

[0056] It is understandable that due to the vehicle changing lanes, turning and other actions during driving, the vehicle does not always drive along the center line. Using only the center line as the target reference line will cause a large deviation in the final result. Therefore, the embodiment of the present application can determine the corresponding target reference line in combination with the vehicle center, the current driving lane and the lane width.

[0057] First, the embodiment of the present application may assume that the lane centerline is the target reference line, and calculate the lateral offset ΔL between the center of the vehicle and the lane centerline of the current driving lane.

[0058] If ΔL>lane width (default 3.5m)*η, the embodiment of the present application can readjust the reference line according to the change direction of ΔL (negative to the left, positive to the right), and use the center line of the adjacent lane on the left or right as the new reference line. η is an empirical coefficient.

[0059] The preset offset threshold may be set by a person skilled in the art according to actual conditions, and is not specifically limited here. The preset width threshold is obtained by the lane width and an empirical coefficient.

[0060] Optionally, in one embodiment of the present application, the longitudinal predicted trajectory of each vehicle is obtained based on the historical fitting trajectory and the change in trajectory points, including: obtaining the displacement of each vehicle in each preset sampling period and obtaining the corresponding displacement change rate; obtaining the second state matrix of the vehicle based on the uniform speed model and the uniform acceleration model based on the displacement change rate and the preset transition probability; and obtaining the longitudinal predicted trajectory of the vehicle based on the second state matrix and the historical fitting trajectory.

[0061] The prediction of longitudinal trajectories is elaborated here.

[0062] The embodiment of the present application can use two models to perform lateral prediction on the target vehicle, namely, the uniform speed model CV and the uniform acceleration model CA. In the uniform speed model, the vehicle moves forward at a uniform speed. In the uniform acceleration model, the vehicle moves forward at a constant acceleration.

[0063] The historical trajectory points of the vehicle are fitted into a curve, and the displacement S of the vehicle in each sampling period is calculated. w , and monitor S w The rate of change of ΔS w .

[0064] If ΔS w If the threshold value has not been exceeded for a period of time, the vehicle is considered to be in the CV state, and the state matrix is ​​reset to [CV, CA] = [0.9, 0.1];

[0065] If ΔS w When the threshold is exceeded, the transition probabilities of the four gradients are set according to the magnitude of the threshold, such as P V-A ∈(1, 0.8, 0.6, 0.4). Update CV and CA values, CV = CV*P V-A , CA = 1-CV;

[0066] When the CV value is lower than 0.5, the state matrix is ​​inverted [CV, CA] = [0.1, 0.9];

[0067] The embodiment of the present application can predict two trajectory point sets for the vehicle, one is the trajectory point set 1 of uniform speed driving, and the other is the trajectory point set 2 of uniform acceleration driving. According to the state matrix [CV, CA], the final predicted trajectory point set is the result of the fusion of trajectory point set 1 and trajectory point set 2. The higher the value of CV, the closer the final predicted trajectory point is to the point on trajectory point set 1, and vice versa.

[0068] Optionally, in one embodiment of the present application, after obtaining the lateral predicted trajectory and longitudinal predicted trajectory of each vehicle based on the historical fitting trajectory and the change in trajectory points, it also includes: screening out all lane-departing vehicles that meet a preset cross-border risk based on the lateral predicted trajectory and the current driving lane; and correcting the lateral predicted trajectory of the lane-departing vehicle based on the lane boundary of the current driving lane.

[0069] During the actual execution process, the vehicle may have a predicted trajectory that exceeds the lane boundary. The embodiment of the present application can correct the predicted trajectory when the predicted trajectory exceeds the boundary of the drivable area such as the curb.

[0070] In step S103, at least two collision risk vehicles with collision risk are screened out based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle.

[0071] As a possible implementation method, the embodiment of the present application can simulate the trajectory of each vehicle according to the lateral predicted trajectory and longitudinal predicted trajectory of each vehicle, and filter out vehicles whose trajectories intersect at the same time, and define the vehicles as collision risk vehicles.

[0072] In step S104, the state matrix stability of each collision risk vehicle is calculated in turn, and the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle are corrected according to the state matrix stability.

[0073] Furthermore, the embodiments of the present application may use the degree of change in the state of the vehicle as a measurement standard, and determine the collision risk vehicles that require lateral predicted trajectory and longitudinal predicted trajectory correction according to the stability of the state matrix of each collision risk vehicle.

[0074] Optionally, in one embodiment of the present application, the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle are corrected according to the state matrix stability, including: setting all collision risk vehicles whose lateral predicted trajectory and longitudinal predicted trajectory intersect at the same position point as a risk collision group; determining the collision risk vehicle with the largest state matrix stability in the risk collision group as the avoidance vehicle, and correcting the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle; judging whether there is a collision risk between the remaining collision risk vehicles based on the lateral predicted trajectory and the longitudinal predicted trajectory of the remaining collision risk vehicles in the risk collision group; if there is a collision risk between the remaining collision risk vehicles, determining the remaining collision risk vehicle with the largest state matrix stability among the remaining collision risk vehicles as the avoidance vehicle, and correcting the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle until there is no collision risk between the remaining collision risk vehicles.

[0075] During the actual implementation process, vehicles with collision risk whose trajectories may intersect at the same location point can be set as a risk collision group, and the number of collision risk vehicles in the risk collision group can be determined. At this time, the embodiment of the present application can perform correction and screening based on the number.

[0076] If the number is 2, the collision risk vehicle with the best state matrix stability can be selected from the two collision risk vehicles to perform trajectory correction to avoid collision;

[0077] If the number is greater than 2, a collision risk vehicle with the best state matrix stability can be selected from multiple collision risk numbers for trajectory correction. After the correction, the collision risk vehicle with the best state matrix stability is selected from the remaining collision risk vehicles and trajectory correction is performed again until there is no collision risk between all collision risk vehicles.

[0078] Among them, the state matrix with the largest stability indicates the best state matrix stability, and whether the state matrix is ​​stable can be determined based on the changes in the state matrix of the corresponding collision risk vehicle over a period of time.

[0079] Optionally, in one embodiment of the present application, after correcting the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle, it also includes: sending the lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding vehicle, so that the corresponding vehicle performs corresponding automatic driving control based on the lateral predicted trajectory and the longitudinal predicted trajectory; sending the corrected lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding collision risk vehicle, so that the corresponding collision risk vehicle performs corresponding automatic driving control based on the corrected lateral predicted trajectory and the longitudinal predicted trajectory.

[0080] After the predicted trajectories of all vehicles are corrected, the embodiment of the present application can send the lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding vehicles so that the vehicles can perform automatic driving control according to the lateral predicted trajectory and the longitudinal predicted trajectory.

[0081] Combination Figures 2 to 5 As shown, the working principle of the vehicle trajectory prediction method of the embodiment of the present application is described in detail with an embodiment.

[0082] like Figure 2 , which is a schematic diagram of the principle of performing lateral prediction of a vehicle using an embodiment of the present application.

[0083] The embodiment of the present application can use two models to predict the lateral direction of the vehicle, namely the straight-moving model S and the turning model T. In the straight-moving model, the vehicle travels along the center line of the current lane, and if it encounters a curve, it still keeps driving in the center of the lane. In the turning model, the vehicle turns to one side according to a constant turning radius.

[0084] First, the embodiment of the present application may assume that the lane centerline is the target reference line, and calculate the lateral offset ΔL between the center of the vehicle and the lane centerline of the current driving lane, continuously track the vehicle through the perception data of the autonomous driving vehicle, and monitor the value of ΔL in each sampling cycle.

[0085] If ΔL does not exceed the threshold value for a period of time, the vehicle is considered to be in the S state and the state matrix is ​​set, such as [S, T] = [0.9, 0.1];

[0086] If ΔL exceeds the threshold, the transition probabilities of the four gradients are set according to the magnitude of the threshold, such as P S-T ∈(1, 0.8, 0.6, 0.4). Update S and T values, S = S*P S-T , T = 1-S;

[0087] When the S value is lower than 0.5, the state matrix is ​​inverted [S, T] = [0.1, 0.9];

[0088] If ΔL>lane width (default 3.5m)*η, the embodiment of the present application can readjust the reference line according to the change direction of ΔL (negative to the left, positive to the right), and use the center line of the adjacent lane on the left or right as the new reference line. η is an empirical coefficient.

[0089] If ΔL<lane width (default 3.5m)*ζ, set the transition probability P of the four gradients according to the deviation of ΔL T-S ∈(1, 0.8, 0.6, 0.4), update T and S values, T = T*P T-S , S=1-T.

[0090] If ΔL ≥ lane width (default 3.5m) * ε and lasts for a period of time, it is considered that the target vehicle is currently in state S, and the state matrix is ​​reset to [S, T] = [0.1, 0.9], where ζ is the empirical coefficient.

[0091] When the T value is lower than 0.5, the state matrix is ​​inverted [S, T] = [0.9, 0.1].

[0092] The embodiment of the present application can fit the historical trajectory points of the vehicle into a curve and predict two trajectories for the vehicle, one is a straight trajectory 1 along the reference line of the target vehicle, and the other is a lane change trajectory 2 with a transverse curvature based on the historical trajectory curve of the vehicle. According to the state matrix [S, T], the final predicted trajectory is the result of the fusion of trajectory 1 and trajectory 2. The higher the value of S, the closer the final predicted trajectory is to trajectory 1, and vice versa.

[0093] like Figure 3 As shown, it is a schematic diagram of the principle of using the embodiment of the present application to predict the longitudinal direction of the vehicle. The embodiment of the present application can use two models to predict the lateral direction of the target vehicle, namely, the uniform speed model CV and the uniform acceleration model CA. Under the uniform speed model, the vehicle moves forward at a constant speed. Under the uniform acceleration model, the vehicle moves forward at a constant acceleration.

[0094] The historical trajectory points of the vehicle are fitted into a curve, and the displacement S of the vehicle in each sampling period is calculated. w , and monitor S w The rate of change of ΔS w .

[0095] If ΔS w If the threshold value has not been exceeded for a period of time, the vehicle is considered to be in the CV state, and the state matrix is ​​reset to [CV, CA] = [0.9, 0.1];

[0096] If ΔS w When the threshold is exceeded, the transition probabilities of the four gradients are set according to the magnitude of the threshold, such as P V-A ∈(1, 0.8, 0.6, 0.4). Update CV and CA values, CV = CV*P V-A , CA = 1-CV;

[0097] When the CV value is lower than 0.5, the state matrix is ​​inverted [CV, CA] = [0.1, 0.9];

[0098] The embodiment of the present application can predict two trajectory point sets for the vehicle, one is the trajectory point set 1 of uniform speed driving, and the other is the trajectory point set 2 of uniform acceleration driving. According to the state matrix [CV, CA], the final predicted trajectory point set is the result of the fusion of trajectory point set 1 and trajectory point set 2. The higher the value of CV, the closer the final predicted trajectory point is to the point on trajectory point set 1, and vice versa.

[0099] like Figure 4 , which is a schematic diagram of the principle of collision detection and avoidance using an embodiment of the present application;

[0100] In order to make the description clearer, target vehicle 1 and target vehicle 2 are used herein to refer to two collision risk vehicles that have a collision risk.

[0101] The embodiment of the present application can perform collision check on the predicted trajectory line, such as Figure 4 As shown in the figure, the trajectory prediction of target vehicle 1 and target vehicle 2 has the possibility of collision, and one of the trajectories needs to be corrected, and the target vehicle avoids target vehicle 2 or the target vehicle 2 avoids target vehicle 1. The degree of state change of the target vehicle is used as a measurement standard, and the specific reference state matrix changes over a period of time. The prediction trajectory of the target vehicle with a more stable state matrix is ​​corrected to avoid collision.

[0102] like Figure 5 , which is a schematic diagram of the principle of performing boundary constraints using an embodiment of the present application.

[0103] The embodiment of the present application can impose boundary constraints on the vehicle trajectory prediction. When the predicted trajectory exceeds the boundary of the drivable area such as the curb, the predicted trajectory needs to be corrected.

[0104] In summary, the embodiments of the present application fully consider the vehicle as an inertial body, and its motion law conforms to the basic mechanical constraints. The state matrix and the transition probability are set, and the state matrix is ​​adjusted or reset according to the changes in the monitored quantity. The results of the trajectory prediction are fused and calculated according to the interactive multi-model algorithm. In addition, to ensure the rationality of the trajectory prediction, based on the detection of the possibility of collision, the state stability of the vehicle is used to evaluate which one is more inclined to avoid the collision. Based on the constraints of the boundary of the drivable area, the predicted trajectory is corrected. The embodiments of the present application can meet the trajectory prediction at a lower computing cost and meet the real-time requirements of the autonomous driving system.

[0105] According to the vehicle trajectory prediction method proposed in the embodiment of the present application, the historical fitting trajectory of each vehicle can be obtained based on the trajectory points uploaded by multiple vehicles, and the lateral predicted trajectory and longitudinal predicted trajectory of each vehicle can be obtained based on the historical fitting trajectory and the change in the trajectory points, thereby screening out at least two collision risk vehicles with collision risks, and correcting the lateral predicted trajectory and longitudinal predicted trajectory of each collision risk vehicle based on the stability of the state matrix of each collision risk vehicle to achieve vehicle trajectory prediction, fully considering that the target vehicle is an inertial body, and its motion law conforms to the basic mechanical constraints. The results of trajectory prediction are fused and calculated according to the interactive multi-model algorithm, and based on the detection of collision possibility, it is evaluated according to the state stability of the vehicle to determine which one is more inclined to avoid collision, and it can cope with environmental changes and sudden road conditions, so as to serve as a reference for the automatic driving control of each vehicle. Thus, the technical problems of poor adaptability to environmental changes, reliance on high-quality driving data, and difficulty in coping with sudden road conditions in the related technologies are solved.

[0106] Next, a vehicle trajectory prediction device according to an embodiment of the present application is described with reference to the accompanying drawings.

[0107] Figure 6 It is a block diagram of a vehicle trajectory prediction device according to an embodiment of the present application.

[0108] like Figure 6 As shown, the vehicle trajectory prediction device 10 is applied to a server, wherein the device 10 includes: a receiving module 100, a prediction module 200, a first screening module 300 and a first correction module 400.

[0109] Specifically, the receiving module 100 is used to receive trajectory points uploaded by multiple vehicles, and obtain the history fitting trajectory of each vehicle based on the trajectory points.

[0110] The prediction module 200 is used to obtain the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle based on the historical matching trajectory and the change amount of the trajectory point.

[0111] The first screening module 300 is used to screen out at least two collision risk vehicles with collision risks based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle.

[0112] The first correction module 400 is used to calculate the state matrix stability of each collision risk vehicle in turn, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the state matrix stability.

[0113] Optionally, in one embodiment of the present application, the prediction module 200 includes: a determination unit, a first calculation unit and a first prediction unit.

[0114] The determination unit is used to determine the corresponding target reference line according to the vehicle center, current driving lane and lane width of each vehicle.

[0115] The first calculation unit is used to obtain a first state matrix based on a straight-moving model and a turning model for each vehicle based on a target reference line, a vehicle center, and a preset transfer probability.

[0116] The first prediction unit is used to obtain a lateral predicted trajectory of the vehicle based on the first state matrix and the historical fitting trajectory of the vehicle.

[0117] Optionally, in one embodiment of the present application, the determination unit includes: a calculation subunit, a first determination subunit and a second determination subunit.

[0118] The calculation subunit is used to calculate the lateral offset between the center of the vehicle and the center line of the lane of the current driving lane.

[0119] The first determination subunit is used to determine that the target reference line is the lane center line when the lateral offset is less than a preset offset threshold.

[0120] The second determination subunit is used to determine whether the lateral offset is greater than a preset width threshold when the lateral offset is greater than or equal to a preset offset threshold. If the lateral offset is greater than the preset width threshold, the target reference line is determined to be the lane center line of the adjacent lane, wherein the preset width threshold is obtained by the lane width.

[0121] Optionally, in one embodiment of the present application, the vehicle trajectory prediction device 10 further includes: a second screening module and a second correction module.

[0122] The second screening module is used to screen out all lane-departing vehicles that meet a preset crossing risk based on the lateral predicted trajectory and the current driving lane.

[0123] The second correction module is used to correct the lateral predicted trajectory of the lane-departing vehicle based on the lane boundary of the current driving lane.

[0124] Optionally, in one embodiment of the present application, the prediction module 200 includes: an acquisition unit, a second calculation unit and a second prediction unit.

[0125] The acquisition unit is used to acquire the displacement of each vehicle in each preset sampling period and obtain the corresponding displacement change rate.

[0126] The second calculation unit is used to obtain a second state matrix of the vehicle based on a uniform speed model and a uniform acceleration model based on a displacement change rate and a preset transfer probability.

[0127] The second prediction unit is used to obtain a predicted longitudinal trajectory of the vehicle based on the second state matrix and the history matching trajectory.

[0128] Optionally, in one embodiment of the present application, the correction module 400 includes: a setting unit, a first correction unit, a judgment unit and a second correction unit.

[0129] The setting unit is used to set all vehicles with collision risks whose lateral predicted trajectories and longitudinal predicted trajectories intersect at the same position point as a risk collision group.

[0130] The first correction unit is used to determine the collision risk vehicle with the largest state matrix stability in the risk collision group as the avoidance vehicle, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle.

[0131] The judgment unit is used to judge whether there is a collision risk between the remaining collision risk vehicles in the risk collision group based on the lateral predicted trajectories and the longitudinal predicted trajectories of the remaining collision risk vehicles.

[0132] The second correction unit is used to determine, when there is a collision risk between the remaining collision risk vehicles, the remaining collision risk vehicle with the largest state matrix stability as the avoidance vehicle, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle until there is no collision risk between the remaining collision risk vehicles.

[0133] Optionally, in one embodiment of the present application, the vehicle trajectory prediction device 10 further includes: a first generating module and a second sending module.

[0134] Among them, the first generating module is used to send the lateral predicted trajectory and the longitudinal predicted trajectory to the corresponding vehicle, so that the corresponding vehicle performs corresponding automatic driving control based on the lateral predicted trajectory and the longitudinal predicted trajectory.

[0135] The second sending module is used to send the corrected lateral predicted trajectory and longitudinal predicted trajectory to the corresponding collision risk vehicle, so that the corresponding collision risk vehicle performs corresponding automatic driving control based on the corrected lateral predicted trajectory and longitudinal predicted trajectory.

[0136] It should be noted that the above explanation of the vehicle trajectory prediction method embodiment is also applicable to the vehicle trajectory prediction device of this embodiment, and will not be repeated here.

[0137] According to the vehicle trajectory prediction device proposed in the embodiment of the present application, the historical fitting trajectory of each vehicle can be obtained based on the trajectory points uploaded by multiple vehicles, and the lateral predicted trajectory and longitudinal predicted trajectory of each vehicle can be obtained based on the historical fitting trajectory and the change in the trajectory points, thereby screening out at least two collision risk vehicles with collision risks, and correcting the lateral predicted trajectory and longitudinal predicted trajectory of each collision risk vehicle based on the stability of the state matrix of each collision risk vehicle to achieve vehicle trajectory prediction, fully considering that the target vehicle is an inertial body, and its motion law conforms to the basic mechanical constraints. The results of trajectory prediction are fused and calculated according to the interactive multi-model algorithm, and based on the detection of collision possibility, it is evaluated according to the state stability of the vehicle to determine which one is more inclined to avoid collision, and it can cope with environmental changes and sudden road conditions, so as to serve as a reference for the automatic driving control of each vehicle. Thus, the technical problems of poor adaptability to environmental changes, reliance on high-quality driving data, and difficulty in coping with sudden road conditions in the related technologies are solved.

[0138] Figure 7 A schematic diagram of the structure of a server provided in an embodiment of the present application. The server may include:

[0139] A memory 701 , a processor 702 , and a computer program stored in the memory 701 and executable on the processor 702 .

[0140] When the processor 702 executes the program, the vehicle trajectory prediction method provided in the above embodiment is implemented.

[0141] Furthermore, the server also includes:

[0142] The communication interface 703 is used for communication between the memory 701 and the processor 702 .

[0143] The memory 701 is used to store computer programs that can be executed on the processor 702 .

[0144] The memory 701 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0145] If the memory 701, the processor 702 and the communication interface 703 are implemented independently, the communication interface 703, the memory 701 and the processor 702 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0146] Optionally, in a specific implementation, if the memory 701, the processor 702 and the communication interface 703 are integrated on a chip, the memory 701, the processor 702 and the communication interface 703 can communicate with each other through an internal interface.

[0147] The processor 702 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0148] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the vehicle trajectory prediction method as described above is implemented.

[0149] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the vehicle trajectory prediction method provided by an embodiment of the present invention.

[0150] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0151] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0152] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0153] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0154] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0155] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0156] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0157] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A vehicle trajectory prediction method, characterized in that: Applied to a server, wherein the method comprises the following steps: Receiving trajectory points uploaded by multiple vehicles, and obtaining a history fitting trajectory of each vehicle based on the trajectory points; Obtaining a lateral predicted trajectory and a longitudinal predicted trajectory of each vehicle based on the history matching trajectory and the amount of change of the trajectory points; Based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle, at least two collision risk vehicles with collision risk are screened out; The state matrix stability of each collision risk vehicle is calculated in turn, and the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle are corrected according to the state matrix stability.

2. The method according to claim 1, characterized in that The step of obtaining the lateral predicted trajectory of each vehicle based on the history matching trajectory and the change amount of the trajectory point includes: Determine a corresponding target reference line according to the vehicle center, the current driving lane and the lane width of each vehicle; Obtaining a first state matrix based on a straight-moving model and a turning model for each vehicle based on the target reference line, the vehicle center, and a preset transition probability; A lateral predicted trajectory of the vehicle is obtained based on the first state matrix and the history fitting trajectory of the vehicle.

3. The method according to claim 2, characterized in that The determining the corresponding target reference line according to the vehicle center, the current driving lane and the lane width of each vehicle includes: Calculating a lateral offset between the center of the vehicle and the center line of the lane of the current driving lane; When the lateral offset is less than a preset offset threshold, determining the target reference line to be the lane center line; When the lateral offset is greater than or equal to the preset offset threshold, determine whether the lateral offset is greater than a preset width threshold; if the lateral offset is greater than the preset width threshold, determine that the target reference line is a lane centerline of an adjacent lane, wherein the preset width threshold is obtained from the lane width.

4. The method according to claim 2, characterized in that: After obtaining the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle based on the history matching trajectory and the variation of the trajectory points, the method further includes: Screening out all lane-departing vehicles that meet a preset crossing risk based on the lateral predicted trajectory and the current driving lane; The predicted lateral trajectory of the lane-departing vehicle is corrected based on the lane boundary of the current driving lane.

5. The method according to claim 1, characterized in that The obtaining of the longitudinal predicted trajectory of each vehicle based on the history matching trajectory and the variation of the trajectory points includes: Obtaining the displacement of each vehicle in each preset sampling period and obtaining the corresponding displacement change rate; Obtaining a second state matrix of the vehicle based on a uniform speed model and a uniform acceleration model based on the displacement change rate and a preset transition probability; A predicted longitudinal trajectory of the vehicle is obtained based on the second state matrix and the history matching trajectory.

6. The method according to claim 1, characterized in that The step of correcting the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the stability of the state matrix includes: All vehicles with collision risks whose lateral predicted trajectories and longitudinal predicted trajectories intersect at the same position point are set as a risk collision group; Determine the collision risk vehicle with the largest state matrix stability in the risk collision group as the avoidance vehicle, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle; Determining whether there is a collision risk between the remaining collision risk vehicles in the risk collision group based on the lateral predicted trajectories and the longitudinal predicted trajectories of the remaining collision risk vehicles; If the collision risk exists between the remaining collision risk vehicles, the remaining collision risk vehicle with the largest state matrix stability among the remaining collision risk vehicles is determined as the avoidance vehicle, and the lateral predicted trajectory and the longitudinal predicted trajectory of the avoidance vehicle are corrected until the collision risk does not exist between the remaining collision risk vehicles.

7. The method according to claim 1, characterized in that After correcting the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle, the method further includes: Sending the lateral predicted trajectory and the longitudinal predicted trajectory to a corresponding vehicle, so that the corresponding vehicle performs corresponding automatic driving control based on the lateral predicted trajectory and the longitudinal predicted trajectory; The corrected lateral predicted trajectory and longitudinal predicted trajectory are sent to the corresponding collision risk vehicle, so that the corresponding collision risk vehicle performs corresponding automatic driving control based on the corrected lateral predicted trajectory and longitudinal predicted trajectory.

8. A vehicle trajectory prediction device, characterized in that: Applied to a server, wherein the device comprises: A receiving module, used for receiving trajectory points uploaded by multiple vehicles, and obtaining a historical fitting trajectory of each vehicle based on the trajectory points; A prediction module, used for obtaining a lateral predicted trajectory and a longitudinal predicted trajectory of each vehicle based on the historical matching trajectory and the variation of the trajectory points; A screening module, for screening out at least two collision risk vehicles with a collision risk based on the lateral predicted trajectory and the longitudinal predicted trajectory of each vehicle; The correction module is used to calculate the state matrix stability of each collision risk vehicle in turn, and correct the lateral predicted trajectory and the longitudinal predicted trajectory of each collision risk vehicle according to the state matrix stability.

9. A server, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the vehicle trajectory prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the vehicle trajectory prediction method as described in any one of claims 1-7.

Images