Road traffic simulation method, device and equipment and storage medium

Abstract

A road traffic simulation method, device and equipment and storage medium are disclosed. The sampling features of map information and the historical trajectory of an obstacle are encoded to obtain a historical position vector representing the historical position of the obstacle in the map. The sampling features of the map information, the future trajectory of the obstacle and the historical position vector are encoded to obtain a posterior distribution vector representing the future position of the obstacle in the map. The state features of the obstacle in a hidden space at a target time are determined based on the historical position vector and the posterior distribution vector. The observable state information of the obstacle at the next time is predicted based on the state features of the obstacle in the hidden space at the target time and the observable state information of all traffic participants including the obstacle at the target time. The posterior information of the future trajectory of the obstacle is introduced in the obstacle prediction model based on the deep neural network, so that the traffic simulation has controllability and intelligence.

Description

Technical Field

[0001] This invention relates to computer technology, and more particularly to a road traffic simulation method, apparatus, equipment, and storage medium. Background Technology

[0002] With the rapid development of technology and the increasing power of computers, real-world traffic can be simulated using computers. Traffic simulation is an important component of intelligent transportation systems and a significant application of computer technology in traffic engineering. It can dynamically and realistically simulate various traffic phenomena such as traffic flow and accidents, reproduce the spatiotemporal changes of traffic flow, and deeply analyze the characteristics of vehicles, drivers, pedestrians, roads, and traffic. This allows for effective research in areas such as traffic planning, traffic organization and management, traffic energy conservation, and the rationalization of material transport flow. Furthermore, through virtual reality technology, traffic simulation can intuitively demonstrate the operation of vehicles on the road network, providing an economical, efficient, and risk-free simulation of whether traffic is congested, whether roads are clear, and whether any accidents have occurred at a specific location.

[0003] Existing simulation models for traffic scenarios mainly fall into two categories: hand-designed models and deep neural network-based models. Hand-designed models require manually designed parameters and only support the simulation of specific types of behavior. While highly controllable, they lack flexibility and are prone to failure in complex interactive scenarios. Deep neural network-based models lack effective means to reproduce scenarios, resulting in a lack of system controllability. Summary of the Invention

[0004] This invention provides a road traffic simulation method, apparatus, equipment, and storage medium, enabling traffic simulation to be both controllable and intelligent, thereby improving the credibility of traffic simulation.

[0005] In a first aspect, the present invention provides a road traffic simulation method, comprising:

[0006] The sampling features of the map information and the historical trajectories of obstacles are encoded to obtain a historical position vector representing the historical position of the obstacle in the map;

[0007] The sampling features of the map information, the future trajectory of the obstacle, and the historical position vector are encoded to obtain the posterior distribution vector representing the future position of the obstacle in the map;

[0008] Based on the historical position vector and the posterior distribution vector, determine the state characteristics of the obstacle in the latent space at the target time;

[0009] Based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time, predict the observable state information of the obstacle at the next time.

[0010] Optionally, before encoding the sampling features of the map information and the historical trajectories of obstacles, the following may also be included:

[0011] The map information is encoded to obtain map encoding features;

[0012] The map encoding information is sampled based on the location of the main vehicle at the target time to obtain the sampling features of the map information.

[0013] Optionally, the sampling features of the map information and the historical trajectories of obstacles are encoded to obtain a historical position vector representing the historical position of the obstacle in the map, including:

[0014] The first concatenated vector is obtained by concatenating the sampled features of the map information with the historical trajectory vector representing the historical trajectory of the obstacle.

[0015] A recurrent neural network is used to process the first stitched vector to obtain a historical position vector representing the historical position of the obstacle in the map.

[0016] Optionally, the sampling features of the map information, the future trajectory of the obstacle, and the historical position vector are encoded to obtain a posterior distribution vector representing the future position of the obstacle in the map, including:

[0017] The sampling features of the map information, the future trajectory vector representing the future trajectory of the obstacle, and the historical position vector are concatenated to obtain the second concatenated vector;

[0018] The second concatenated vector is processed using a recurrent neural network to obtain a posterior distribution vector representing the future position of the obstacle in the map.

[0019] Optionally, determining the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector includes:

[0020] Based on the historical location of the obstacle, the posterior distribution vector is sampled to obtain the sampling features of the posterior distribution vector;

[0021] By fusing the sampling features of the posterior distribution vector and the historical position vector, the state features of the obstacle in the latent space at the target time are obtained.

[0022] Optionally, based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time, the observable state information of the obstacle at the next time is predicted, including:

[0023] By integrating the observable state information of all traffic participants, including the aforementioned obstacles, at the target time, a summary description of the traffic scene at the target time is obtained;

[0024] The observable state information of the obstacle at the target time and the summary description of the traffic scene at the target time are input into the pre-built obstacle interaction behavior model for processing to obtain the interaction characteristics of the obstacle with other traffic participants at the target time.

[0025] The third spliced ​​feature is obtained by splicing together the interaction features of the obstacle with other traffic participants at the target time, the sampling features of map information, and the state features of the obstacle in the latent space at the target time.

[0026] The third splicing feature is input into a pre-built prediction model for processing to obtain the observable state information of the obstacle at the next moment.

[0027] Optionally, the prediction model also outputs the state features of the obstacle in the latent space at the next time step, and the method further includes:

[0028] Determine whether the next moment is a preset moment;

[0029] If not, then the next moment is taken as the target moment, and the process is returned to perform the step of fusing the observable state information of all traffic participants, including the obstacles, at the target moment to obtain a summary description of the traffic scene at the target moment;

[0030] If so, the simulation process ends.

[0031] Secondly, the present invention also provides a road traffic simulation device, comprising:

[0032] The first encoding module is used to encode the sampling features of map information and the historical trajectory of obstacles to obtain a historical position vector representing the historical position of the obstacle in the map;

[0033] The second encoding module is used to encode the sampling features of the map information, the future trajectory of the obstacle, and the historical position vector to obtain the posterior distribution vector representing the future position of the obstacle in the map.

[0034] The state feature determination module is used to determine the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector.

[0035] The state information prediction module is used to predict the observable state information of the obstacle at the next time moment based on the state characteristics of the obstacle in the hidden space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time.

[0036] Thirdly, the present invention also provides an electronic device, comprising:

[0037] One or more processors;

[0038] Memory, used to store one or more programs;

[0039] When the one or more programs are executed by the one or more processors, the one or more processors implement the road traffic simulation method provided in the first aspect of the present invention.

[0040] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the road traffic simulation method provided in the first aspect of the present invention.

[0041] The road traffic simulation method provided by this invention includes: encoding the sampling features of map information and the historical trajectories of obstacles to obtain a historical position vector representing the historical position of the obstacle on the map; encoding the sampling features of map information, the future trajectory of the obstacle, and the historical position vector to obtain a posterior distribution vector representing the future position of the obstacle on the map; determining the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector; and predicting the observable state information of the obstacle at the next time step based on the state features of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time. This invention introduces posterior information of the future trajectory of obstacles into the obstacle prediction model based on deep neural networks, making the model output controllable, thereby enabling traffic simulation to be both controllable and intelligent, and improving the credibility of traffic simulation. Furthermore, by extracting the unobservable state features of obstacles in the latent space, this invention can realistically simulate the interaction between obstacles and the main vehicle and other obstacles, improving the intelligence and credibility of traffic simulation.

[0042] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 A flowchart of a road traffic simulation method provided in an embodiment of the present invention;

[0045] Figure 2 A flowchart for extracting state features provided in an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram illustrating the prediction of the next state information of an obstacle, provided by an embodiment of the present invention.

[0047] Figure 4 This is a schematic diagram of the structure of a road traffic simulation device provided in an embodiment of the present invention;

[0048] Figure 5 This is a schematic diagram of the structure of an electronic device provided as an embodiment of the present invention.

[0049] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0050] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0051] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0052] Figure 1 This is a flowchart illustrating a road traffic simulation method provided in an embodiment of the present invention. This embodiment is applicable to road traffic simulation. The method can be executed by a road traffic simulation device provided in this embodiment of the present invention. This device can be implemented in software and / or hardware, and is typically configured in an electronic device, such as... Figure 1 As shown, the road traffic simulation method specifically includes the following steps:

[0053] S101. Encode the sampling features of the map information and the historical trajectory of the obstacle to obtain the historical position vector representing the historical position of the obstacle in the map.

[0054] Map information can be pre-constructed high-precision map information. High-precision maps refer to navigation maps with high resolution and high abundance of features, where both absolute and relative accuracy are at the centimeter level (10 to 20 centimeters). They include the lowest-level static high-precision map and other dynamic information. The static high-precision map contains lane models, road components, road attributes, and other positioning layers. The lane model contains detailed road information, such as lane lines, lane center lines, and lane attribute changes. Furthermore, the lane model also needs to include mathematical parameters such as road curvature, slope, heading, and cross slope. Autonomous driving dynamic information refers to all dynamic information within the intelligent connected system, generally including map dynamic information, sensor information, driving behavior, and traffic dynamic information management.

[0055] The sampling features of the map information can be features within a preset range for the main vehicle, which is the autonomous driving vehicle. The historical trajectory of the obstacle can be the set of obstacle positions on the map before the target time. In this embodiment of the invention, the obstacle can be a traffic participant such as a vehicle, pedestrian, or non-motorized vehicle, and this embodiment of the invention is not limited thereto. For example, the target time can be the initial time at which the simulation begins.

[0056] Figure 2 A flowchart for extracting state features is provided in an embodiment of the present invention, such as... Figure 2 As shown, in this embodiment of the invention, the sampling features (map t) of map information and the historical trajectory (history traj) of obstacles are encoded to obtain a historical position vector (history v) representing the historical position of obstacles in the map. This historical position vector (history v) not only represents the historical position information of obstacles, but also anchors to the map information.

[0057] In some embodiments of the present invention, such as Figure 2 As shown, a Convolutional Neural Network (CNN) can be used to encode map information (map fea) to obtain map encoded features (map emb). For example, the CNN can include U-net, etc., and this embodiment of the invention is not limited thereto. In other embodiments of the invention, a graph neural network (e.g., Vector Net) can also be used to encode map information. In other embodiments of the invention, an attention mechanism (e.g., Scene Transformer) can also be used to encode map information, and this embodiment of the invention is not limited thereto. After obtaining the map encoded features (map emb), the map encoded features are sampled based on the position of the main vehicle at the target time to obtain the sampled features of the map information (map t). For example, based on the position of the main vehicle at the target time, features within a preset range of that position are collected from the map encoded features as the sampled features of the map information (map t). Specifically, as... Figure 2 As shown, the location of the obstacle at the target time and the map encoding features (map emb) are input into a pre-constructed feature extraction network (birdview) for processing to obtain the sampled features of the map information (map t). The feature extraction network (birdview) can use the nearest neighbor bilinear interpolation algorithm to process the map encoding features, and this embodiment of the invention is not limited thereto.

[0058] In some embodiments of the present invention, such as Figure 2 As shown, the sampled features of the map information (map t) are concatenated with the historical trajectory vector (history traj) representing the historical trajectory of the obstacle to obtain the first concatenated vector. Then, a recurrent neural network (RNN) is used to process the first concatenated vector to obtain the historical position vector (history v) representing the historical position of the obstacle in the map. For example, as shown... Figure 2As shown, a recurrent neural network can be a Long Short-Term Memory (LSTM) network. LSTM is a special type of recurrent neural network, mainly designed to solve the gradient vanishing and gradient exploding problems during long sequence training.

[0059] S102. Encode the sampling features of the map information, the future trajectory of the obstacle, and the historical position vector to obtain the posterior distribution vector representing the future position of the obstacle in the map.

[0060] In some embodiments of the present invention, a time-recurrent neural network model is used to encode the sampling features of map information, the future trajectories of obstacles, and historical position vectors to obtain a posterior distribution vector representing the future position of obstacles on the map. The posterior distribution is a probability distribution obtained using probability theory to obtain the conditional probability distribution based on the sample (future trajectories of obstacles) distribution and the prior distribution of unknown parameters. The future trajectories of obstacles are the set of positions of obstacles after the target time (this data is known). The embodiments of the present invention introduce posterior information about obstacle trajectories into the obstacle prediction model based on deep neural networks, making the model output controllable, thereby enabling traffic simulation to be both controllable and intelligent.

[0061] For example, such as Figure 2 As shown, the sampled features of the map information (map t), the future trajectory vector representing the future trajectory of the obstacle (future traj), and the historical position vector (history v) are concatenated to obtain a second concatenated vector. Then, a recurrent neural network is used to process the second concatenated vector to obtain the posterior distribution vector representing the future position of the obstacle on the map. For example, as shown... Figure 2 As shown, the recurrent neural network can be a long short-term memory network (LSTM).

[0062] S103. Determine the state characteristics of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector.

[0063] In this embodiment of the invention, the state features (state{*, t}) of the obstacle in the latent space at the target time are determined based on the historical position vector and the posterior distribution vector, where * is the obstacle's number and t is the target time. The state features of the obstacle in the latent space represent unobservable states, such as the obstacle's intention.

[0064] Exemplary examples, in some embodiments of the present invention, such as Figure 2As shown, the posterior distribution vector is sampled based on the historical position of the obstacle to obtain the sampled features (posterior t) of the posterior distribution vector. Then, the sampled features (posterior t) of the posterior distribution vector and the historical position vector (history v) are fused to obtain the state features (state{*, t}) representing the obstacle in the latent space at the target time.

[0065] S104. Based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time, predict the observable state information of the obstacle at the next time.

[0066] After obtaining the state characteristics of the obstacle in the latent space at the target time, the observable state information of the obstacle at the next time is predicted based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time.

[0067] Figure 3 This is a schematic diagram illustrating the prediction of the next moment's state information of an obstacle, provided as an embodiment of the present invention. Exemplarily, in some embodiments of the present invention, step S104 includes the following sub-steps:

[0068] 1. By integrating the observable state information of all traffic participants, including obstacles, at the target time, a summary description of the traffic scene at the target time is obtained.

[0069] For example, such as Figure 3 As shown, taking the observable state information of an obstacle (agent 0) at the next moment as an example, the observable state information (y{*, t}) of all traffic participants (including the obstacle and the ego car) at the target moment is fused and summarized to obtain a summary description st of the traffic scene at the target moment. For example, the observable state information may include the displacement of the traffic participants relative to the previous moment, their speed, orientation, etc., which are not limited in this embodiment of the invention.

[0070] 2. Input the observable state information of the obstacle at the target time and the summary description of the traffic scene at the target time into the pre-built obstacle interaction behavior model for processing, and obtain the interaction characteristics of the obstacle with other traffic participants at the target time.

[0071] For example, such as Figure 3As shown, the observable state information (y{0, t}) of the obstacle (agent 0) at the target time and the summarized description st of the traffic scene at the target time are input into a pre-built obstacle interaction behavior model (interact) for processing, thereby obtaining the interaction features of the obstacle (agent 0) with other traffic participants at the target time. For example, the obstacle interaction behavior model (interact) may include a Transformer model, a PointNet model, an MCG block model, etc., and this embodiment of the invention is not limited thereto.

[0072] 3. The interaction features of the obstacle with other traffic participants at the target time, the sampling features of map information, and the state features of the obstacle in the latent space at the target time are spliced ​​together to obtain the third spliced ​​feature.

[0073] For example, such as Figure 3 As shown, the interaction features of the obstacle (agent 0) with other traffic participants at the target time, the sampling features of map information (map t), and the state features (state{0, t}) of the obstacle (agent 0) in the latent space at the target time are concatenated to obtain the third concatenated feature.

[0074] 4. Input the third splicing feature into the pre-built prediction model for processing to obtain the observable state information of the obstacle at the next moment.

[0075] For example, such as Figure 3 As shown, the third concatenated feature is input into a pre-built prediction model (step cell) for processing to obtain the observable state information (y{0,t+1}) of the obstacle (agent 0) at the next time step and the state features (state{0,t+1}) in the latent space. For example, the prediction model (step cell) can be a long short-term memory network model or other time-recurrent network models, which is not limited in this embodiment of the invention.

[0076] The prediction process for the state information of other obstacles is similar to that for obstacle (agent 0), and will not be described again in this embodiment of the invention.

[0077] For example, after obtaining the state information of the obstacle at the next moment, it is determined whether the next moment is a preset moment, where the preset moment can be the time when the simulation ends. If not, the next moment is taken as the target moment, and the process returns to perform the step of fusing the state information of all traffic participants, including the obstacle, at the target moment to obtain a summary description of the traffic scene at the target moment. Taking the obstacle (agent 0) as an example, this involves fusing the observable state information (y{*, t+1}) of all traffic participants predicted in the previous step at time t+1, inputting the observable state information of the obstacle (agent 0) at time t+1 and the summary description of the traffic scene at time t+1 into a pre-built obstacle interaction behavior model for processing, and obtaining the interaction characteristics of the obstacle (agent 0) with other traffic participants at time t+1. The interaction features of the obstacle (agent 0) with other traffic participants at time t+1, the sampling features of map information, and the state features of the obstacle in the latent space at time t+1 (state{0, t+1}) are concatenated. This concatenated feature is then input into a pre-built prediction model for processing, yielding the state information (y{0, t+2}) of the obstacle (agent 0) at time t+2 and its state features in the latent space (state{0, t+2}). If the result is satisfactory, the simulation process ends. In this way, the state information of each obstacle at every time after the start of the simulation is obtained.

[0078] The road traffic simulation method provided in this invention includes: encoding the sampling features of map information and the historical trajectories of obstacles to obtain a historical position vector representing the historical position of the obstacle in the map; encoding the sampling features of map information, the future trajectory of the obstacle, and the historical position vector to obtain a posterior distribution vector representing the future position of the obstacle in the map; determining the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector; and predicting the observable state information of the obstacle at the next time moment based on the state features of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time. This invention introduces posterior information of the future trajectory of obstacles into the obstacle prediction model based on deep neural networks, making the model output controllable, thereby enabling traffic simulation to be both controllable and intelligent, and improving the credibility of traffic simulation. Furthermore, by extracting the unobservable state features of obstacles in the latent space, this invention can realistically simulate the interaction between obstacles and the main vehicle and other obstacles, improving the intelligence and credibility of traffic simulation.

[0079] This invention also provides a road traffic simulation device. Figure 4 This is a schematic diagram of the structure of a road traffic simulation device provided in an embodiment of the present invention, as shown below. Figure 4 As shown, the road traffic simulation device includes:

[0080] The first encoding module 201 is used to encode the sampling features of map information and the historical trajectory of obstacles to obtain a historical position vector representing the historical position of the obstacle in the map;

[0081] The second encoding module 202 is used to encode the sampling features of the map information, the future trajectory of the obstacle, and the historical position vector to obtain the posterior distribution vector representing the future position of the obstacle in the map;

[0082] The state feature determination module 203 is used to determine the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector.

[0083] The state information prediction module 204 is used to predict the observable state information of the obstacle at the next time moment based on the state characteristics of the obstacle in the hidden space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time.

[0084] In some embodiments of the present invention, the road traffic simulation device further includes:

[0085] The map encoding module is used to encode the map information before encoding the sampled features of the map information and the historical trajectories of obstacles, so as to obtain the map encoding features;

[0086] The map sampling module is used to sample the map encoding information based on the location of the main vehicle at the target time to obtain the sampling features of the map information.

[0087] In some embodiments of the present invention, the first encoding module 201 includes:

[0088] The first stitching unit is used to stitch together the sampled features of the map information with the historical trajectory vector representing the historical trajectory of the obstacle to obtain the first stitching vector;

[0089] The first processing unit is used to process the first spliced ​​vector using a recurrent neural network to obtain a historical position vector representing the historical position of the obstacle in the map.

[0090] In some embodiments of the present invention, the second encoding module 202 includes:

[0091] The second stitching unit is used to stitch together the sampling features of the map information, the future trajectory vector representing the future trajectory of the obstacle, and the historical position vector to obtain the second stitching vector;

[0092] The second processing unit is used to process the second spliced ​​vector using a recurrent neural network to obtain a posterior distribution vector representing the future position of the obstacle in the map.

[0093] In some embodiments of the present invention, the state feature determination module 203 includes:

[0094] A posterior sampling unit is used to sample the posterior distribution vector based on the historical position of the obstacle to obtain the sampling features of the posterior distribution vector;

[0095] The fusion unit is used to fuse the sampling features of the posterior distribution vector and the historical position vector to obtain the state features of the obstacle in the latent space at the target time.

[0096] In some embodiments of the present invention, the state information prediction module 204 includes:

[0097] The aggregation unit is used to integrate the observable state information of all traffic participants, including the obstacles, at the target time to obtain a summary description of the traffic scene at the target time.

[0098] The interaction feature extraction unit is used to input the observable state information of the obstacle at the target time and the summary description of the traffic scene at the target time into the pre-built obstacle interaction behavior model for processing, so as to obtain the interaction features of the obstacle with other traffic participants at the target time.

[0099] The third stitching unit is used to stitch together the interaction features of the obstacle with other traffic participants at the target time, the sampling features of map information, and the state features of the obstacle in the latent space at the target time to obtain the third stitching feature.

[0100] The prediction unit is used to input the third splicing feature into a pre-built prediction model for processing to obtain the observable state information of the obstacle at the next moment.

[0101] In some embodiments of the present invention, the prediction model also outputs the state characteristics of the obstacle in the latent space at the next moment, and the road traffic simulation device further includes:

[0102] The judgment module is used to determine whether the next moment is a preset moment;

[0103] The execution module is used to take the next time as the target time when the next time is not the preset time, and return to the execution module to perform the step of fusing the observable state information of all traffic participants, including the obstacles, at the target time to obtain a summary description of the traffic scene at the target time.

[0104] The termination module is used to end the simulation process when the next time step is a preset time step.

[0105] The road traffic simulation device described above can execute the road traffic simulation method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the road traffic simulation method.

[0106] This application provides an electronic device. Figure 5 This is a schematic diagram of an electronic device provided for an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0107] like Figure 5 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0108] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0109] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as road traffic simulation methods.

[0110] In some embodiments, the road traffic simulation method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the road traffic simulation method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the road traffic simulation method by any other suitable means (e.g., by means of firmware).

[0111] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0112] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0113] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0114] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0115] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0116] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0117] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the road traffic simulation method provided in any embodiment of this application.

[0118] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0119] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0120] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A road traffic simulation method, characterized in that, include: The sampling features of the map information and the historical trajectories of obstacles are encoded to obtain a historical position vector representing the historical position of the obstacle in the map; The sampling features of the map information, the future trajectory of the obstacle, and the historical position vector are encoded to obtain the posterior distribution vector representing the future position of the obstacle in the map; Based on the historical position vector and the posterior distribution vector, determine the state characteristics of the obstacle in the latent space at the target time; Based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time, predict the observable state information of the obstacle at the next time. Based on the state characteristics of the obstacle in the latent space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time, predict the observable state information of the obstacle at the next time, including: By integrating the observable state information of all traffic participants, including the aforementioned obstacles, at the target time, a summary description of the traffic scene at the target time is obtained; The observable state information of the obstacle at the target time and the summary description of the traffic scene at the target time are input into the pre-built obstacle interaction behavior model for processing to obtain the interaction characteristics of the obstacle with other traffic participants at the target time. The third spliced ​​feature is obtained by splicing together the interaction features of the obstacle with other traffic participants at the target time, the sampling features of map information, and the state features of the obstacle in the latent space at the target time. The third splicing feature is input into a pre-built prediction model for processing to obtain the observable state information of the obstacle at the next moment.

2. The road traffic simulation method according to claim 1, characterized in that, Before encoding the sampling features of map information and the historical trajectories of obstacles, the following steps are also included: The map information is encoded to obtain map encoding features; The map encoding information is sampled based on the location of the main vehicle at the target time to obtain the sampling features of the map information.

3. The road traffic simulation method according to claim 1, characterized in that, Encoding the sampling features of map information and the historical trajectories of obstacles yields a historical position vector representing the historical location of the obstacle on the map, including: The first concatenated vector is obtained by concatenating the sampled features of the map information with the historical trajectory vector representing the historical trajectory of the obstacle. A recurrent neural network is used to process the first stitched vector to obtain a historical position vector representing the historical position of the obstacle in the map.

4. The road traffic simulation method according to any one of claims 1-3, characterized in that, Encoding the sampling features of the map information, the future trajectories of obstacles, and the historical position vectors yields a posterior distribution vector representing the future position of the obstacles on the map, including: The sampling features of the map information, the future trajectory vector representing the future trajectory of the obstacle, and the historical position vector are concatenated to obtain the second concatenated vector; The second concatenated vector is processed using a recurrent neural network to obtain a posterior distribution vector representing the future position of the obstacle in the map.

5. The road traffic simulation method according to any one of claims 1-3, characterized in that, Determining the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector includes: Based on the historical location of the obstacle, the posterior distribution vector is sampled to obtain the sampling features of the posterior distribution vector; By fusing the sampling features of the posterior distribution vector and the historical position vector, the state features of the obstacle in the latent space at the target time are obtained.

6. The road traffic simulation method according to claim 1, characterized in that, The prediction model also outputs the state features of the obstacle in the latent space at the next time step, and the method further includes: Determine whether the next moment is a preset moment; If not, then the next moment is taken as the target moment, and the process is returned to perform the step of fusing the observable state information of all traffic participants, including the obstacles, at the target moment to obtain a summary description of the traffic scene at the target moment; If so, the simulation process ends.

7. A road traffic simulation device, characterized in that, The method for performing the road traffic simulation method according to any one of claims 1-6 includes: The first encoding module is used to encode the sampling features of map information and the historical trajectory of obstacles to obtain a historical position vector representing the historical position of the obstacle in the map; The second encoding module is used to encode the sampling features of the map information, the future trajectory of the obstacle, and the historical position vector to obtain the posterior distribution vector representing the future position of the obstacle in the map. The state feature determination module is used to determine the state features of the obstacle in the latent space at the target time based on the historical position vector and the posterior distribution vector. The state information prediction module is used to predict the observable state information of the obstacle at the next time moment based on the state characteristics of the obstacle in the hidden space at the target time and the observable state information of all traffic participants, including the obstacle, at the target time.

8. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the road traffic simulation method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the road traffic simulation method as described in any one of claims 1-6.

Images