Traffic participant trajectory prediction method and system based on multiple interactive behaviors

Abstract

The present invention provides a method and system for predicting the trajectory of traffic participants based on multiple interactive behaviors, including: obtaining high-precision map data and node features of traffic participants, and constructing a static interaction layer and a dynamic interaction layer; dynamically interacting the static interaction layer and the dynamic interaction layer based on a preset traffic light information gated neural network to construct an interactive network; performing supervised learning training on the interactive network through a preset target loss function, outputting a trajectory prediction network model, and performing trajectory prediction through the trajectory prediction network model. The present invention solves the problem of inaccurate trajectory prediction of traffic participants in the prior art, realizes the mutual fusion of multiple traffic interactive behaviors, and completes accurate trajectory prediction.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation technology, and in particular to a method and system for predicting the trajectory of traffic participants based on multiple interactive behaviors. Background Art

[0002] The future trajectory prediction of traffic participants is to predict the future driving trajectory of traffic participants such as pedestrians and vehicles based on their current or historical trajectory and environmental information, so that the autonomous driving vehicle can make advance decisions based on the prediction results and appropriately adjust the trajectory when driving along the planned path to avoid collision with the traffic participants and drive safely to the destination.

[0003] The current method of predicting the future trajectory of traffic participants does not take into account the mutual influence of the motion states of traffic participants in real scenarios, that is, social interaction. It only predicts the future trajectory of each traffic participant based on the historical trajectory point position coordinates of each independent individual by treating each traffic participant as an independent individual. Since social interaction will also affect the running trajectory of traffic participants, collisions and interferences are prone to occur when making driving decisions based on the prediction results obtained by this prediction method. For self-driving cars, there are more vehicle trajectory prediction technologies for specific scenarios or highway scenarios, while there are relatively few trajectory prediction technologies for traffic participants on open urban roads. Moreover, most technologies often ignore the interactive behaviors between traffic participants and the constraints of roads and traffic rules. Summary of the invention

[0004] The present invention provides a method and system for predicting the trajectory of traffic participants based on multiple interactive behaviors, which are used to solve the problem of inaccurate prediction of the trajectory of traffic participants in the prior art, realize the mutual fusion of multiple traffic interactive behaviors, and complete the accurate prediction of the trajectory.

[0005] The present invention provides a method for predicting the trajectory of traffic participants based on multiple interactive behaviors, comprising:

[0006] Obtain high-precision map data and node features of traffic participants, and construct static interaction layers and dynamic interaction layers;

[0007] Based on a preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network;

[0008] The interactive network is trained through supervised learning using a preset target loss function, a trajectory prediction network model is output, and trajectory prediction is performed using the trajectory prediction network model.

[0009] According to a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention, the high-precision map data and node features of traffic participants are obtained, and a static interaction layer and a dynamic interaction layer are constructed, including:

[0010] Performing vectorized feature extraction on the high-precision map data, constructing road node features, and adding road type information;

[0011] The key points in the road node features are feature encoded to complete the construction of the static interaction layer.

[0012] According to a traffic participant trajectory prediction method based on multiple interactive behaviors provided by the present invention, vectorized feature extraction is performed on the high-precision map data, road node features are constructed, and road type information is added, which specifically includes:

[0013] The road node features include: road structure features and stop line structure features, selecting the starting point and direction of the road centerline, and evenly extracting key points from the line within the same spatial distance;

[0014] Adjacent key points are connected in sequence, and road type information is added to complete the construction of road structure features and stop line structure features.

[0015] According to a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention, the high-precision map data and node features of traffic participants are obtained, and a static interaction layer and a dynamic interaction layer are constructed, which also includes:

[0016] The traffic participant detects surrounding traffic participants through a preset perception system;

[0017] According to the detection results, each traffic participant is regarded as a node, and the node features are encoded using a linear layer;

[0018] The encoded node features are used to establish interactions between dynamic layers through a global attention mechanism and processed through sparse logistic regression to complete the construction of the dynamic interaction layer.

[0019] According to a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention, the preset traffic light information gated neural network dynamically interacts the static interaction layer and the dynamic interaction layer to construct an interactive network, specifically including:

[0020] Encoding the information of the traffic light, and after encoding, performing feature representation of the traffic light through a linear layer;

[0021] The key point features in the static interaction layer, the node features in the dynamic interaction layer, and the traffic light features are fused through a gated neural network to obtain implicit features.

[0022] The implicit features are processed by a preset sigmoid nonlinear function to obtain a dynamic and static interaction weight to control the interaction strength;

[0023] Based on the interaction strength, an interaction relationship between traffic participants and road nodes is established through a local graphic attention mechanism, local attention feature aggregation is performed, and an interaction network is established.

[0024] According to a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention, the interactive network is subjected to supervised learning training through a preset target loss function, a trajectory prediction network model is output, and trajectory prediction is performed through the trajectory prediction network model, specifically comprising:

[0025] The interaction network is supervised by the target loss function;

[0026] The target loss function includes the RMSE loss between the predicted estimate and the target trajectory, the KLD loss between the predicted trajectory and the target trajectory, and outputs the trajectory prediction network model after training is completed;

[0027] The node features of traffic participants in the dynamic interaction layer are input into the trajectory decoder in the trajectory prediction network model to decode and predict the future trajectory of each traffic participant.

[0028] The present invention also provides a traffic participant trajectory prediction system based on multiple interactive behaviors, the system comprising:

[0029] The interactive layer construction module is used to obtain high-precision map data and node features of traffic participants, and to construct static and dynamic interactive layers;

[0030] An interactive network construction module, used to dynamically interact the static interactive layer and the dynamic interactive layer based on a preset traffic light information gating neural network to construct an interactive network;

[0031] The prediction module is used to perform supervised learning training on the interactive network through a preset target loss function, output a trajectory prediction network model, and perform trajectory prediction through the trajectory prediction network model.

[0032] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for predicting the trajectory of traffic participants based on multiple interactive behaviors as described above is implemented.

[0033] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for predicting the trajectory of traffic participants based on multiple interactive behaviors as described in any one of the above is implemented.

[0034] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned methods for predicting the trajectory of traffic participants based on multiple interactive behaviors.

[0035] The present invention provides a method and system for predicting the trajectory of traffic participants based on multiple interactive behaviors. By acquiring high-precision map data and node features of traffic participants, a static interaction layer and a dynamic interaction layer are constructed. The static interaction layer and the dynamic interaction layer are dynamically interacted based on a preset traffic light information gated neural network to construct an interactive network. The interactive network is supervised and trained through a preset target loss function, and a trajectory prediction network model is output. Trajectory prediction is performed through the trajectory prediction network model. Information in a large batch of scene data is learned, and big data prior knowledge is further utilized in the process of static layer construction, thereby achieving better prediction results than the existing technology and achieving accurate trajectory prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 It is one of the flow charts of a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention;

[0038] Figure 2 This is the second flow chart of a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention;

[0039] Figure 3 This is a third flow chart of a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention;

[0040] Figure 4 This is a fourth flow chart of a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention;

[0041] Figure 5 This is a fifth flow chart of a method for predicting the trajectory of traffic participants based on multiple interactive behaviors provided by the present invention;

[0042] Figure 6 It is a schematic diagram of module connection of a traffic participant trajectory prediction system based on multiple interactive behaviors provided by the present invention;

[0043] Figure 7 It is a schematic diagram of performing sparse-softmax processing on a dynamic interaction graph provided by the present invention;

[0044] Figure 8 It is a structural schematic diagram of the electronic device provided by the present invention.

[0045] Reference numerals:

[0046] 110: Interaction layer building module; 120: Interaction network building module; 130: Prediction module;

[0047] 810: processor; 820: communication interface; 830: memory; 840: communication bus. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] Combine the following Figure 1-Figure 5 A method for predicting the trajectory of traffic participants based on multiple interactive behaviors of the present invention is described, comprising:

[0050] S100, obtain high-precision map data and node features of traffic participants, and construct static interaction layer and dynamic interaction layer;

[0051] S200, dynamically interacting the static interaction layer and the dynamic interaction layer based on a preset traffic light information gated neural network to construct an interaction network;

[0052] S300, performing supervised learning training on the interactive network through a preset target loss function, outputting a trajectory prediction network model, and performing trajectory prediction through the trajectory prediction network model.

[0053] The present invention can construct dynamic and static complex urban traffic scene interaction diagrams through high-precision maps and autonomous driving perception systems. The node features in the dynamic and static interaction layers are encoded through neural networks, and the interaction relationship between dynamic and static nodes is carried out through gated neural networks using traffic light information, thereby realizing complex interaction relationship modeling; based on the interaction relationship network and trajectory prediction decoder, the future trajectory of traffic participants can be predicted. This technology realizes the full modeling of traffic participant information and the accurate prediction of their future trajectory, which can provide accurate decision-making prior information for autonomous driving vehicles.

[0054] Obtain high-precision map data and node features of traffic participants, and construct static interaction layers and dynamic interaction layers, including:

[0055] S101, performing vectorized feature extraction on the high-precision map data, constructing road node features, and adding road type information;

[0056] S102: feature encoding the key points in the road node features to complete the construction of the static interaction layer.

[0057] Vectorized feature extraction is performed on the high-precision map data to construct road node features and add road type information, specifically including:

[0058] S1011, the road node features include: road structure features and stop line structure features, selecting the starting point and direction of the road centerline, and evenly extracting key points from the line within the same spatial distance;

[0059] S1012: Connect adjacent key points in sequence and add road type information to complete the construction of road structure features and stop line structure features.

[0060] In the present invention, vectorized feature extraction is performed on the high-precision map of the city's roads. For the construction of road structural features, a starting point and direction of a road centerline are selected, key points are uniformly extracted from the spline within the same spatial distance, and adjacent key points are connected in sequence. The key point information includes the position, the curvature of the road at this point, and the direction vector of the connection point with the predecessor; the structural feature construction of the stop line is similar, except that the feature construction process of the key point will add road type information. Finally, the linear layer is used to encode the key points in the static interaction layer.

[0061] Obtain high-precision map data and node features of traffic participants, and construct static and dynamic interaction layers, including:

[0062] The traffic participant detects surrounding traffic participants through a preset perception system;

[0063] According to the detection results, each traffic participant is regarded as a node, and the node features are encoded using a linear layer;

[0064] The encoded node features are used to establish interactions between dynamic layers through a global attention mechanism and processed through sparse logistic regression to complete the construction of the dynamic interaction layer.

[0065] Before constructing the dynamic interaction layer of traffic participants, the perception system of the autonomous vehicle is first used to detect the traffic participants around the vehicle. Each traffic participant can be used as a node in the dynamic interaction layer. The node features are composed of the position, speed, acceleration, and type of the traffic participant, and the node features are encoded using a linear layer. At the same time, in order to preliminarily establish the interaction relationship between the traffic participants in the dynamic structure graph, the Global graph attention mechanism is used to establish the interaction between the dynamic layers. However, since the Global graph attention mechanism allows each node to participate in the attention mechanism of any other node, this mechanism will cause the phenomenon of "over-interaction". For this reason, after completing the global attention of the dynamic interaction graph, sparse-softmax (sparse logistic regression) is used for further processing to screen out the interaction pairs whose interaction weights are lower than a certain threshold. Reference Figure 7 Schematic diagram of the dynamic interaction layer before and after processing.

[0066] Based on the preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network, which specifically includes:

[0067] S201, encoding the information of the traffic light, and after encoding, performing feature representation of the traffic light through a linear layer;

[0068] S202, fusing the key point features in the static interaction layer, the node features in the dynamic interaction layer, and the traffic light features through a gated neural network to obtain implicit features;

[0069] S203, processing the implicit features with a preset sigmoid nonlinear function to obtain a dynamic-static interaction weight to control the interaction strength;

[0070] S204: Based on the interaction strength, an interaction relationship between traffic participants and road nodes is established through a local graphic attention mechanism, local attention feature aggregation is performed, and an interaction network is established.

[0071] In the present invention, traffic light information can be understood as a connecting medium between traffic participants and roads. The change of traffic light status controls the open and closed status of roads, as well as the open and closed status of traffic participants. Based on this mechanism, it can be considered that traffic lights are an important factor in controlling the dynamic layer and the static layer. Therefore, the information of traffic lights is used as a gate switch to control the interaction strength of the dynamic and static interaction layers.

[0072] First, the traffic light information is encoded, and the light state information, phase information, and light countdown information of the traffic light are input into the linear layer for feature representation. Secondly, the feature D of each node in the dynamic layer is extracted. f For the static layer, the central discrete node of the lane where each traffic participant is located is encoded into lane-level feature S f The dynamic interaction layer representative features and the static interaction layer representative features are combined with the traffic light features T f The gate mechanism (gated neural network) is integrated to obtain the implicit features of the traffic lights controlling the roads and traffic participants. Finally, the implicit features are passed through the sigmoid nonlinear function to obtain the dynamic and static interaction weights (0-1) to control the interaction intensity.

[0073] G f =sigmoid(Attention(T f , S f , D f )

[0074] G f When it approaches 0, it means that the lane is in a closed state. For traffic participants, the traffic behavior is single and tends to slow down and stop. Therefore, the interaction with the road in the static interaction layer is relatively weak, and the interaction with the stop line in the static interaction layer is relatively strong. On the contrary, G f A larger value indicates that the traffic participant is likely to travel along the road, which requires stronger interaction. Under the control of the interaction strength, the Local Graph Attention mechanism is used to establish the interaction relationship between the traffic participant and the road node. According to the location of the traffic participant, the road nodes in the four directions closest to it (east, west, south, and north in geographical orientation) are selected in the static interaction layer to perform local attention feature aggregation. The final interaction method is as follows: fL is the static layer lane node feature, S fS Stop the line node feature for static layers.

[0075]

[0076] After the interactive network is established, the characteristics of traffic participants in each dynamic layer are sent to the trajectory decoder to decode the future trajectory of each traffic participant.

[0077] The interactive network is subjected to supervised learning training through a preset target loss function, and a trajectory prediction network model is outputted. Trajectory prediction is performed through the trajectory prediction network model, specifically including:

[0078] S301, supervise the interaction network through the target loss function;

[0079] S302 The target loss function includes the RMSE loss between the predicted estimate and the target trajectory, the KLD loss between the predicted trajectory and the target trajectory, and outputs the trajectory prediction network model after the training is completed;

[0080] S303, inputting the node features of the traffic participants in the dynamic interaction layer into the trajectory decoder in the trajectory prediction network model, decoding and predicting the future trajectory of each traffic participant.

[0081] The above interaction network is supervised by the target loss function, which includes the RMSE loss between the predicted estimate and the target trajectory, and the KLD loss between the predicted trajectory and the target trajectory. After the above training process, a traffic participant trajectory prediction network based on multiple interactive behaviors is obtained. The trajectory is predicted by the trajectory prediction network.

[0082] When the present invention is applied to actual scenarios, the future behavior of traffic participants can be accurately predicted through high-precision maps and perception information. This feature enables the present invention to be applied in complex urban road scenarios.

[0083] In addition, the present invention utilizes the currently developed and mature deep learning method to learn the information in large quantities of scene data, and further utilizes the prior knowledge of big data in the process of static layer construction, thereby achieving better prediction results than the existing technology. Since the present invention does not require the support of hardware equipment and only needs to be processed in the algorithm, it can be directly applied to existing vehicle-road collaboration, vehicle networking and other systems, and has strong portability.

[0084] refer to Figure 6 The present invention also discloses a traffic participant trajectory prediction system based on multiple interactive behaviors, the system comprising:

[0085] The interactive layer construction module 110 is used to obtain high-precision map data and node features of traffic participants, and to construct a static interactive layer and a dynamic interactive layer;

[0086] An interactive network construction module 120, configured to dynamically interact the static interactive layer with the dynamic interactive layer based on a preset traffic light information gated neural network to construct an interactive network;

[0087] The prediction module 130 is used to perform supervised learning training on the interactive network through a preset target loss function, output a trajectory prediction network model, and perform trajectory prediction through the trajectory prediction network model.

[0088] The interactive layer construction module extracts vectorized features from the high-precision map data, constructs road node features, and adds road type information;

[0089] The key points in the road node features are feature encoded to complete the construction of the static interaction layer.

[0090] Vectorized feature extraction is performed on the high-precision map data to construct road node features and add road type information, specifically including:

[0091] The road node features include: structural features and stop line structural features, selecting the starting point and direction of the road centerline, and evenly extracting key points from the line within the same spatial distance;

[0092] Adjacent key points are connected in sequence, and road type information is added to complete the construction of road structure features and stop line structure features.

[0093] The traffic participant detects surrounding traffic participants through a preset perception system;

[0094] According to the detection results, each traffic participant is regarded as a node, and the node features are encoded using a linear layer;

[0095] The encoded node features are used to establish interactions between dynamic layers through a global attention mechanism and processed through sparse logistic regression to complete the construction of the dynamic interaction layer.

[0096] An interactive network building module encodes the information of the traffic light and represents the features of the traffic light through a linear layer after encoding;

[0097] The key point features in the static interaction layer, the node features in the dynamic interaction layer, and the traffic light features are fused through a gated neural network to obtain implicit features.

[0098] The implicit features are processed by a preset sigmoid nonlinear function to obtain a dynamic and static interaction weight to control the interaction strength;

[0099] Based on the interaction strength, an interaction relationship between traffic participants and road nodes is established through a local graphic attention mechanism, local attention feature aggregation is performed, and an interaction network is established.

[0100] The prediction module supervises the interaction network through the target loss function;

[0101] The target loss function includes the RMSE loss between the predicted estimate and the target trajectory, the KLD loss between the predicted trajectory and the target trajectory, and outputs the trajectory prediction network model after training is completed;

[0102] The node features of traffic participants in the dynamic interaction layer are input into the trajectory decoder in the trajectory prediction network model to decode and predict the future trajectory of each traffic participant.

[0103] The present invention provides a traffic participant trajectory prediction system based on multiple interactive behaviors. By acquiring high-precision map data and node features of traffic participants, a static interaction layer and a dynamic interaction layer are constructed. The static interaction layer and the dynamic interaction layer are dynamically interacted based on a preset traffic light information gated neural network to construct an interactive network. The interactive network is supervised and trained through a preset target loss function, and a trajectory prediction network model is output. Trajectory prediction is performed through the trajectory prediction network model. Information in a large batch of scene data is learned, and big data prior knowledge is further utilized in the static layer construction process, thereby achieving better prediction results than the existing technology and achieving accurate trajectory prediction.

[0104] Figure 8 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 8 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute a method for predicting the trajectory of traffic participants based on multiple interactive behaviors, the method comprising: obtaining high-precision map data and node features of traffic participants, and constructing a static interaction layer and a dynamic interaction layer;

[0105] Based on a preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network;

[0106] The interactive network is trained through supervised learning using a preset target loss function, a trajectory prediction network model is output, and trajectory prediction is performed using the trajectory prediction network model.

[0107] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0108] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute a traffic participant trajectory prediction method based on multiple interactive behaviors provided by the above methods, the method comprising: obtaining high-precision map data and node features of traffic participants, and constructing a static interaction layer and a dynamic interaction layer;

[0109] Based on a preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network;

[0110] The interactive network is trained through supervised learning using a preset target loss function, a trajectory prediction network model is output, and trajectory prediction is performed using the trajectory prediction network model.

[0111] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute a method for predicting the trajectory of a traffic participant based on multiple interactive behaviors provided by the above methods, the method comprising: obtaining high-precision map data and node features of traffic participants, and constructing a static interaction layer and a dynamic interaction layer;

[0112] Based on a preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network;

[0113] The interactive network is trained through supervised learning using a preset target loss function, a trajectory prediction network model is output, and trajectory prediction is performed using the trajectory prediction network model.

[0114] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0115] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the trajectory of traffic participants based on multiple interactive behaviors, characterized in that: include: Obtain high-precision map data and node features of traffic participants, and construct static interaction layers and dynamic interaction layers; Based on a preset traffic light information gated neural network, the static interaction layer and the dynamic interaction layer are dynamically interacted to construct an interaction network; Performing supervised learning training on the interactive network through a preset target loss function, outputting a trajectory prediction network model, and performing trajectory prediction through the trajectory prediction network model; The preset traffic light information gated neural network dynamically interacts the static interaction layer with the dynamic interaction layer to construct an interaction network, specifically including: Encoding the information of the traffic light, and after encoding, performing feature representation of the traffic light through a linear layer; Extract the features of each node in the dynamic layer For the static layer, the central discrete node of the lane where each traffic participant is located is encoded into lane-level features , the dynamic interaction layer representative features and the static interaction layer representative features are combined with the traffic light features The gated neural network is integrated to obtain the implicit features of the traffic lights controlling the roads and traffic participants. Finally, the implicit features are passed through the sigmoid nonlinear function to obtain the dynamic and static interaction weights. , used to control the interaction strength; ； Under the control of interaction intensity, the local graph attention mechanism is used to establish the interaction relationship between traffic participants and road nodes. According to the location of traffic participants, the road nodes in the four nearest directions are selected in the static interaction layer to perform local attention feature aggregation. The final interaction mode is: ； in is the static layer driving lane node feature, Stop the line node feature for static layers.

2. The method for predicting the trajectory of traffic participants based on multiple interactive behaviors according to claim 1 is characterized in that: The acquisition of high-precision map data and node features of traffic participants and the construction of static interaction layer and dynamic interaction layer include: Performing vectorized feature extraction on the high-precision map data, constructing road node features, and adding road type information; The key points in the road node features are feature encoded to complete the construction of the static interaction layer.

3. The method for predicting the trajectory of traffic participants based on multiple interactive behaviors according to claim 2 is characterized in that: Vectorized feature extraction is performed on the high-precision map data to construct road node features and add road type information, specifically including: The road node features include: road structure features and stop line structure features, selecting the starting point and direction of the road centerline, and evenly extracting key points from the line within the same spatial distance; Adjacent key points are connected in sequence, and road type information is added to complete the construction of road structure features and stop line structure features.

4. The method for predicting the trajectory of traffic participants based on multiple interactive behaviors according to claim 1, characterized in that: The obtaining of high-precision map data and node features of traffic participants and the construction of a static interaction layer and a dynamic interaction layer also includes: The traffic participant detects surrounding traffic participants through a preset perception system; According to the detection results, each traffic participant is regarded as a node, and the node features are encoded using a linear layer; The encoded node features are used to establish interactions between dynamic layers through a global attention mechanism and processed through sparse logistic regression to complete the construction of the dynamic interaction layer.

5. The method for predicting the trajectory of traffic participants based on multiple interactive behaviors according to claim 1, characterized in that: The interactive network is subjected to supervised learning training through a preset target loss function, and a trajectory prediction network model is outputted. Trajectory prediction is performed through the trajectory prediction network model, specifically including: The interaction network is supervised by the target loss function; The target loss function includes the RMSE loss between the predicted estimate and the target trajectory, the KLD loss between the predicted trajectory and the target trajectory, and outputs the trajectory prediction network model after training is completed; The node features of traffic participants in the dynamic interaction layer are input into the trajectory decoder in the trajectory prediction network model to decode and predict the future trajectory of each traffic participant.

6. A traffic participant trajectory prediction system based on multiple interactive behaviors, characterized in that: The system comprises: The interactive layer construction module is used to obtain high-precision map data and node features of traffic participants, and to construct static and dynamic interactive layers; An interactive network construction module, used to dynamically interact the static interactive layer and the dynamic interactive layer based on a preset traffic light information gated neural network to construct an interactive network; A prediction module, used to perform supervised learning training on the interactive network through a preset target loss function, output a trajectory prediction network model, and perform trajectory prediction through the trajectory prediction network model; The preset traffic light information gated neural network dynamically interacts the static interaction layer with the dynamic interaction layer to construct an interaction network, specifically including: Encoding the information of the traffic light, and after encoding, performing feature representation of the traffic light through a linear layer; Extract the features of each node in the dynamic layer For the static layer, the central discrete node of the lane where each traffic participant is located is encoded into lane-level features , the dynamic interaction layer representative features and the static interaction layer representative features are combined with the traffic light features The gated neural network is integrated to obtain the implicit features of the traffic lights controlling the roads and traffic participants. Finally, the implicit features are passed through the sigmoid nonlinear function to obtain the dynamic and static interaction weights. , used to control the interaction strength; ； Under the control of interaction intensity, the local graph attention mechanism is used to establish the interaction relationship between traffic participants and road nodes. According to the location of traffic participants, the road nodes in the four nearest directions are selected in the static interaction layer to perform local attention feature aggregation. The final interaction mode is: ； in is the static layer driving lane node feature, Stop the line node feature for static layers.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for predicting the trajectory of traffic participants based on multiple interactive behaviors as described in any one of claims 1 to 5 is implemented.

8. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the trajectory of traffic participants based on multiple interactive behaviors as described in any one of claims 1 to 5 is implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting the trajectory of traffic participants based on multiple interactive behaviors as described in any one of claims 1 to 5 is implemented.

Images