Trajectory fusion methods, devices, equipment, media and products

Abstract

This application provides a trajectory fusion method, apparatus, device, medium, and product, relating to the field of intelligent transportation technology. The method includes: acquiring the current vehicle trajectory of at least one vehicle identified by a sensing device; if a vehicle is detected to be traveling into a blind zone between the sensing devices, determining that a vehicle trajectory prediction condition or a vehicle trajectory simulation condition is met; when the vehicle trajectory prediction condition is met, predicting the vehicle trajectory and fusing the obtained current predicted vehicle trajectory with the current vehicle trajectory of the corresponding vehicle to obtain a current first fused trajectory; when the vehicle trajectory simulation condition is met, performing a microscopic simulation of the vehicle trajectory and fusing the obtained current simulated vehicle trajectory with the current vehicle trajectory of the corresponding vehicle to obtain a current second fused trajectory. By predicting or microscopically simulating the vehicle trajectory, the vehicle trajectory when traveling to the blind zone can be completed, forming a complete vehicle trajectory.

Description

Technical Field

[0001] This application relates to the field of intelligent transportation technology, and in particular to a trajectory fusion method, apparatus, equipment, medium and product. Background Technology

[0002] Vehicle-road cooperation utilizes advanced wireless communication and next-generation Internet technologies to implement comprehensive, dynamic, real-time information exchange between vehicles and infrastructure. Based on the collection and fusion of dynamic traffic information across all times and spaces, it conducts active vehicle safety control and cooperative road management, fully realizing effective coordination among people, vehicles, and roads to ensure traffic safety, improve traffic efficiency, and thus form a safe, efficient, and environmentally friendly road traffic system.

[0003] Vehicle-road cooperation is a core component of smart highway construction. The application of vehicle-road cooperation in smart highways requires the realization of vehicle perception across the entire roadside and the formation of complete vehicle trajectories.

[0004] Full-segment vehicle perception along the roadside relies on a large number of sensing devices such as video cameras and sensors. However, the current distribution density of sensing devices on highways does not meet the requirements for full-segment vehicle perception. There are blind spots between sensing devices, making it impossible to achieve full coverage and realize full-segment vehicle perception. Furthermore, it is impossible to form a complete vehicle trajectory based on the perceived vehicle trajectory. Summary of the Invention

[0005] This application provides a trajectory fusion method, apparatus, device, medium, and product to solve the problems in the prior art where the distribution density of sensing devices on highways does not meet the requirements for vehicle perception across the entire road segment, blind spots exist between sensing devices, and full coverage cannot be achieved, thus failing to realize vehicle perception across the entire road segment.

[0006] Firstly, this application provides a trajectory fusion method, including:

[0007] Acquire the current vehicle trajectory of at least one vehicle identified by the sensing device;

[0008] If the vehicle is detected to be traveling into a blind zone between the sensing devices, then the vehicle trajectory prediction condition or the vehicle trajectory simulation condition is determined to be met.

[0009] When the vehicle trajectory prediction conditions are met, the vehicle trajectory is predicted, and the obtained current vehicle prediction trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory.

[0010] When the vehicle trajectory simulation conditions are met, a microscopic simulation of the vehicle trajectory is performed, and the obtained current vehicle simulation trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

[0011] Secondly, this application provides a trajectory fusion device, comprising:

[0012] The acquisition module is used to acquire the current vehicle movement trajectory of at least one vehicle identified by the sensing device;

[0013] If the module detects that the vehicle is traveling into a blind zone between the sensing devices, it determines that the vehicle trajectory prediction conditions or vehicle trajectory simulation conditions are met.

[0014] The prediction module predicts the vehicle trajectory when the vehicle trajectory prediction conditions are met.

[0015] The fusion module is used to fuse the obtained current vehicle predicted trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory;

[0016] The simulation module is used to perform microscopic simulation of the vehicle trajectory when the vehicle trajectory simulation conditions are met;

[0017] The fusion module is further configured to fuse the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

[0018] Thirdly, this application provides an electronic device, including: a processor, and a memory and a transceiver communicatively connected to the processor;

[0019] The memory stores computer-executed instructions; the transceiver is used for sending and receiving data.

[0020] The processor executes computer execution instructions stored in the memory to implement the method as described in the first aspect.

[0021] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the method described in the first aspect.

[0022] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the method described in the first aspect.

[0023] This application provides a trajectory fusion method, apparatus, device, medium, and product, which acquires the current vehicle motion trajectory of at least one vehicle identified by a sensing device; if the vehicle is detected to be traveling into a blind zone between the sensing devices, it is determined that a vehicle trajectory prediction condition or a vehicle trajectory simulation condition is met; when the vehicle trajectory prediction condition is met, the vehicle trajectory is predicted, and the obtained current predicted vehicle trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain a current first fused trajectory; when the vehicle trajectory simulation condition is met, the vehicle trajectory is microscopically simulated, and the obtained current simulated vehicle trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain a current second fused trajectory. By predicting the vehicle trajectory or performing microscopic simulation of the vehicle trajectory, a fused trajectory can be formed. The fused trajectory is used to complete the vehicle trajectory when the vehicle travels to the blind zone position, forming a complete vehicle trajectory; and by determining that the vehicle trajectory prediction condition or the vehicle trajectory simulation condition is met, a suitable method for completing the vehicle trajectory can be selected, which can improve the accuracy of the completed vehicle trajectory. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0025] Figure 1 This is a schematic diagram of the high-speed smart pole erection for this application;

[0026] Figure 2 A flowchart of a trajectory fusion method provided in Embodiment 1 of this application;

[0027] Figure 3 This is a schematic diagram of the sensing device structure provided in an embodiment of this application;

[0028] Figure 4 This is a schematic diagram of the trajectory fusion method provided in the embodiments of this application;

[0029] Figure 5 This is a diagram illustrating how vehicle trajectories are sent to a digital display screen.

[0030] Figure 6 This is a schematic diagram showing the vehicle trajectory displayed on a large digital screen.

[0031] Figure 7 This is a schematic diagram illustrating the transmission of vehicle trajectory data to the onboard terminal.

[0032] Figure 8 This is a schematic diagram showing the complete vehicle trajectory displayed on the in-vehicle terminal.

[0033] Figure 9 This is a schematic diagram illustrating a method for predicting vehicle trajectories.

[0034] Figure 10 This is a flowchart of a trajectory fusion method provided in Embodiment 4 of this application;

[0035] Figure 11 This is a schematic diagram of the trajectory fusion device provided in Embodiment Six of this application;

[0036] Figure 12 This is a schematic diagram of the structure of the electronic device provided in Embodiment 7 of this application.

[0037] 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

[0038] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.

[0039] The terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. In the following descriptions of embodiments, "a plurality of" means two or more, unless otherwise explicitly defined.

[0040] To clearly understand the technical solution of this application, the solutions of the prior art will be described in detail first.

[0041] In existing technologies, the method for forming vehicle trajectories involves acquiring vehicle perception data through sensing devices and generating vehicle motion trajectories based on this data, then merging adjacent vehicle motion trajectories to form a complete vehicle trajectory. Currently, sensing devices are typically installed on highway smart poles. Due to the limitations of the installation location of these poles, the distance between two poles may exceed the sensing range of the devices, creating blind spots. This results in the sensing devices' distribution density on highways not meeting the requirements for full-segment vehicle perception, making it impossible to achieve full-segment vehicle perception. When a vehicle travels to a blind spot, its motion trajectory cannot be acquired, causing trajectory interruptions and preventing the formation of a complete vehicle trajectory.

[0042] Figure 1 This is a schematic diagram of the installation of smart poles on high-speed highways. Figure 1 The system includes a high-speed smart pole 1 and a sensing device 2 mounted on the high-speed smart pole 1. The sensing device 2 may include millimeter-wave radar, lidar, high-definition cameras, etc. Figure 1 The sensing range of the middle sensing device 2 is 50 meters to 200 meters, such as Figure 1 As shown, if the adjacent high-speed smart pole 1 is erected at a height of more than 250 meters, there will be blind spots between the sensing devices, and the sensing range of sensing device 2 cannot achieve full coverage.

[0043] Because blind spots exist between sensing devices, it is impossible to perceive vehicles across the entire road segment, thus preventing the formation of complete vehicle trajectories. Therefore, trajectory prediction or microscopic simulation can be used to supplement the vehicle's trajectory when it reaches the blind spot, thereby forming a complete vehicle trajectory. To ensure the accuracy of supplementing the vehicle's trajectory when it reaches the blind spot, trajectory prediction can be performed when the blind spot is small, such as less than 100 meters. However, when the blind spot expands to a certain extent, such as exceeding 100 meters, the error in the predicted vehicle trajectory will accumulate, leading to collisions and other conflicts. Therefore, microscopic simulation is needed to simulate the vehicle trajectory to avoid conflicts caused by excessive trajectory errors.

[0044] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0045] Example 1

[0046] Figure 2 This is a flowchart of a trajectory fusion method provided in Embodiment 1 of this application. This embodiment addresses the problem in existing technologies where the distribution density of sensing devices on highways is insufficient to meet the requirements for full-segment vehicle perception, resulting in blind spots between sensing devices and an inability to achieve full coverage and vehicle perception across the entire road segment. A trajectory fusion method is provided to address this issue. The execution entity of the trajectory fusion method provided in this embodiment can be a trajectory fusion device. In practical applications, this trajectory fusion device can be implemented through computer programs, such as application software, or through media storing relevant computer programs, such as USB flash drives, optical discs, or the cloud. Alternatively, it can be implemented through physical or virtual devices that integrate or install relevant computer programs, such as chips or circuit boards.

[0047] Furthermore, the trajectory fusion device can be located in an electronic device. This electronic device can be a digital computer representing various forms, such as cloud servers, cellular phones, smartphones, laptops, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers.

[0048] like Figure 2As shown, the trajectory fusion method provided in this embodiment includes the following steps:

[0049] Step S101: Obtain the current vehicle movement trajectory of at least one vehicle identified by the sensing device.

[0050] The current vehicle trajectory is the vehicle's driving trajectory determined by the sensing device based on the currently collected sensing information.

[0051] Specifically, the sensing device collects vehicle perception information and synthesizes it into a corresponding vehicle motion trajectory. This trajectory is then sent to a trajectory fusion device, enabling the device to acquire the current vehicle motion trajectory of at least one vehicle identified by the sensing device. The vehicle perception information includes image information, location information, lane information, speed information, etc., collected by the sensing device.

[0052] Figure 3 This is a schematic diagram of the sensing device structure provided in an embodiment of this application. The sensing device is a device capable of collecting vehicle sensing information, such as... Figure 3 As shown, the sensing device includes at least one sensing unit and a processing unit. The sensing unit is used to collect sensing information of vehicles within a certain road segment, and the processing unit is used to synthesize the vehicle's motion trajectory based on the vehicle's sensing information. After collecting the sensing information, the sensing unit sends the sensing information to the processing unit, enabling the processing unit to synthesize the vehicle's motion trajectory based on the vehicle's sensing information. The sensing unit may include at least one of the following: millimeter-wave radar, lidar, video sensing equipment, etc.; the processing unit may be a terminal box; and the video sensing equipment may be a high-definition camera.

[0053] Step S102: If a blind zone between the vehicle and the sensing device is detected, then the vehicle trajectory prediction condition or vehicle trajectory simulation condition is met.

[0054] It should be understood that sensing devices can only collect sensing information about vehicles on road sections within their collection range. If the collection ranges of adjacent sensing devices do not overlap or connect, blind spots will exist between the sensing devices. The trajectory fusion device can pre-observe the collection range of each sensing device.

[0055] Specifically, the current vehicle trajectory can be used to determine whether the vehicle has entered the blind zone between the sensing devices. If the vehicle has entered the blind zone between the sensing devices, the conditions for vehicle trajectory prediction or vehicle trajectory simulation can be determined based on the blind zone between the sensing devices.

[0056] The embodiments of this application do not limit the method of determining whether a vehicle has driven into the blind spot between sensing devices.

[0057] For example, one possible implementation of determining whether a vehicle has driven into a blind zone between sensing devices is as follows: determine whether the vehicle has reached the edge of the sensing range of the sensing device; if the vehicle has reached the edge of the sensing range of the sensing device, determine whether the vehicle's current position is within the sensing range of an adjacent sensing device; if the vehicle's current position is not within the sensing range of an adjacent sensing device, then it can be determined that the vehicle has driven into a blind zone between the sensing devices.

[0058] Another alternative implementation for determining whether a vehicle has entered the blind zone between sensing devices is as follows: pre-determine the blind zone between sensing devices based on the collection range of each sensing device, determine the distance between the current vehicle position and the blind zone, and determine whether the vehicle will enter the blind zone between sensing devices based on the distance between the current vehicle position and the blind zone.

[0059] In this embodiment of the application, if the size of the blind spot range to which the vehicle travels is within the vehicle trajectory prediction capability, then the vehicle trajectory prediction condition is met, and step S103 is executed to predict the vehicle trajectory; if the size of the blind spot range to which the vehicle travels exceeds the vehicle trajectory prediction capability, then the vehicle trajectory simulation condition is met, and step S104 is executed to perform microscopic simulation of the vehicle trajectory.

[0060] This application does not limit the method of determining whether vehicle trajectory prediction conditions or vehicle trajectory simulation conditions are met. For example, a blind zone range corresponding to the vehicle trajectory prediction capability can be preset. If the blind zone range the vehicle travels into is within the preset range, then the vehicle trajectory prediction condition is met; if the blind zone range the vehicle travels into is not within the preset range, then the vehicle trajectory simulation condition is met. Alternatively, a first preset blind zone range threshold can be set. By determining whether the blind zone range the vehicle travels into is greater than the first preset threshold, it is determined whether the vehicle trajectory prediction condition or the vehicle trajectory simulation condition is met.

[0061] Step S103: When the vehicle trajectory prediction conditions are met, predict the vehicle trajectory and fuse the obtained current vehicle prediction trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory.

[0062] Among them, the current vehicle predicted trajectory is the vehicle trajectory predicted based on the current vehicle movement trajectory, and the current first fused trajectory is the vehicle trajectory obtained by fusing the current vehicle movement trajectory and the current vehicle predicted trajectory.

[0063] In this embodiment, when the vehicle trajectory prediction conditions are met, the vehicle trajectory is predicted within a certain future time period by combining historical vehicle trajectories and utilizing the vehicle's most recent trajectory. For example, historical vehicle trajectories can be used as training samples to train the vehicle trajectory prediction model, and the vehicle trajectory within a future time period can be predicted based on the trained model. Any suitable neural network model can be used as the vehicle trajectory prediction model.

[0064] In this embodiment, the method of fusing the obtained current vehicle predicted trajectory with the current vehicle motion trajectory of the corresponding vehicle in step S103, and the method of fusing the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle in step S104, are similar to the method of fusing vehicle motion trajectories obtained through different sensing devices. They all determine two vehicle trajectories that are connected or overlap in time for the same vehicle and fuse the two vehicle trajectories. The specific methods will not be described in detail here.

[0065] Step S104: When the vehicle trajectory simulation conditions are met, perform microscopic simulation of the vehicle trajectory, and fuse the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

[0066] Among them, the current vehicle simulation trajectory is the vehicle trajectory obtained by performing microscopic simulation of the vehicle trajectory, and the first fused trajectory is the vehicle trajectory obtained by fusing the current vehicle motion trajectory with the current vehicle simulation trajectory.

[0067] In this embodiment, when the vehicle trajectory simulation conditions are met, road network information can be combined to set the vehicle's current position, current speed, and vehicle type, and the vehicle simulation trajectory can be obtained using path planning and trajectory generation methods. For example, a microscopic simulation model can be used to perform microscopic simulation of the vehicle trajectory. The microscopic simulation model takes each vehicle as the research object and simulates the driving behavior of each vehicle to obtain the current vehicle simulation trajectory.

[0068] Figure 4This is a schematic diagram of the trajectory fusion method provided in an embodiment of this application. The trajectory fusion device may include a fusion module, a prediction module, and a simulation module. The trajectory fusion device can acquire the vehicle motion trajectory of at least one vehicle identified by the sensing device; the prediction module can preload a vehicle trajectory prediction model and use the vehicle trajectory prediction model to predict the vehicle trajectory in real time to form a predicted vehicle trajectory; the simulation module can preload a high-precision road network, and after acquiring the vehicle motion trajectory, it can add vehicles corresponding to the vehicle motion trajectory, process the departure of the vehicle, and update the vehicle simulation trajectory in the high-precision road network in real time; the fusion module can preload the high-precision road network, acquire the vehicle motion trajectory, the vehicle prediction trajectory, and the vehicle simulation trajectory, and fuse the vehicle motion trajectory and the vehicle prediction trajectory of the same vehicle on the high-precision road network, or fuse the vehicle motion trajectory and the vehicle simulation trajectory to form the complete vehicle trajectory of the vehicle.

[0069] The trajectory fusion method provided in this embodiment acquires the current vehicle movement trajectory of at least one vehicle identified by a sensing device. If a vehicle is detected to be traveling into a blind zone between the sensing devices, it is determined that a vehicle trajectory prediction condition or a vehicle trajectory simulation condition is met. When the vehicle trajectory prediction condition is met, the vehicle trajectory is predicted, and the obtained current predicted vehicle trajectory is fused with the current vehicle movement trajectory of the corresponding vehicle to obtain a current first fused trajectory. When the vehicle trajectory simulation condition is met, the vehicle trajectory is microscopically simulated, and the obtained current simulated vehicle trajectory is fused with the current vehicle movement trajectory of the corresponding vehicle to obtain a current second fused trajectory. By predicting the vehicle trajectory or performing microscopic simulation of the vehicle trajectory, a fused trajectory can be formed. The fused trajectory is used to complete the vehicle trajectory when the vehicle travels to the blind zone position, forming a complete vehicle trajectory. Furthermore, by determining whether the vehicle trajectory prediction condition or the vehicle trajectory simulation condition is met, a suitable method for completing the vehicle trajectory can be selected, which can improve the accuracy of the completed vehicle trajectory.

[0070] Optionally, based on the above embodiments, if step S102 detects a blind zone between the vehicle and the sensing device, then determining whether the vehicle trajectory prediction conditions or vehicle trajectory simulation conditions are met, an optional implementation includes the following steps:

[0071] Step S1021: If a blind spot between the vehicle and the sensing device is detected, the blind spot range is determined.

[0072] In this embodiment of the application, the specific locations of both ends of all blind zones can be determined in advance, such as the station numbers corresponding to both ends of the blind zone. The specific locations of both ends of the blind zone determine the blind zone range of the blind zone. The specific locations of both ends of the blind zone and the blind zone range are associated and stored in any database.

[0073] Specifically, after detecting that the vehicle has entered a blind spot between the sensing devices, the blind spot range of the blind spot can be obtained from the database based on the vehicle's current location.

[0074] Step S1022: If the blind zone range is determined to be less than or equal to the first preset blind zone range threshold, then the vehicle trajectory prediction condition is determined to be met.

[0075] Step S1023: If the blind zone range is determined to be greater than the first preset blind zone range threshold, then the vehicle trajectory simulation conditions are determined to be met.

[0076] In the implementation of this application, a first preset blind zone range threshold can be preset according to the vehicle trajectory prediction capability. For example, if the predicted vehicle trajectory exceeds 100m, the accuracy of the predicted vehicle trajectory beyond 100m will drop significantly, so the first preset blind zone range threshold can be set to 100m.

[0077] For example, the first preset blind zone range threshold is 100m. If the blind zone between the vehicle and the sensing device is less than or equal to 100m, the vehicle trajectory prediction condition is met, and the vehicle trajectory is predicted. If the blind zone between the vehicle and the sensing device is greater than 100m, the vehicle trajectory simulation condition is met, and the vehicle trajectory is microscopically simulated.

[0078] The trajectory fusion method provided in this embodiment can quickly determine whether the blind zone range is greater than the first preset blind zone range threshold, thereby determining whether the vehicle trajectory prediction conditions or vehicle trajectory simulation conditions are met. A suitable method for completing the vehicle trajectory can be selected, which can improve the accuracy of the completed vehicle trajectory.

[0079] Optionally, in this embodiment, the trajectory fusion device can be located on a cloud server. After obtaining the current first fused trajectory or the current second fused trajectory, the device can also send the first fused trajectory or the current second fused trajectory to a digital parallel world, such as a digital screen, so that the current first fused trajectory or the current second fused trajectory can be displayed in the digital parallel world.

[0080] Figure 5 A diagram illustrating the transmission of vehicle trajectories to a digital display screen, such as... Figure 5As shown, all sensing devices can send the current vehicle trajectory of at least one vehicle to a message queue in the cloud server. The cloud server performs full trajectory fusion or prediction, microscopic simulation, and other processing on the vehicle trajectory in the message queue according to the first-in-first-out principle to form the complete vehicle trajectory of all vehicles corresponding to the current vehicle trajectory. The complete vehicle trajectory of the vehicle includes the first fused trajectory or obtains the current second fused trajectory. The complete vehicle trajectory of all vehicles corresponding to the current vehicle trajectory is then sent to the digital screen, so that the complete vehicle trajectory of the vehicle is displayed on the digital screen.

[0081] Figure 6 This is a schematic diagram of the vehicle trajectory displayed on the digital screen, such as... Figure 6 As shown, a panoramic view of all vehicles in the road network can be displayed in real time based on the current movement trajectories of all vehicles, and video data collected by a specific sensing device can be displayed in the upper right corner. The data of all vehicles in the road network displayed in the panoramic view is synchronized with the video data collected by the sensing device.

[0082] Optionally, in this embodiment, the electronic device can be an edge device, and the trajectory fusion device can be in the edge device, which can be a server distributed along the highway. After obtaining the current first fused trajectory or the current second fused trajectory, the first fused trajectory or the current second fused trajectory can be sent to the vehicle terminal so that the first fused trajectory or the current second fused trajectory is displayed in the vehicle terminal.

[0083] Figure 7 This is a diagram illustrating the transmission of vehicle trajectory data to the onboard terminal, as shown below. Figure 7As shown, the information can be transmitted to the vehicle-to-everything (V2X) communication link. The Roadside Unit (RSU) is the roadside unit in the V2X link, and the Onboard Unit (OBU) is the onboard unit in the V2X link. Specifically, the edge device can obtain the current vehicle trajectory of at least one vehicle identified by its corresponding sensing device, such as the sensing device between the edge device and the next edge device. Following a first-in-first-out (FIFO) principle, the edge device performs regional trajectory fusion or prediction, microscopic simulation, and other processing on the vehicle trajectories in the message queue on the server to form a complete vehicle trajectory for the vehicle traveling on the corresponding road segment. The complete vehicle trajectory includes a first fused trajectory or a currently obtained second fused trajectory. This complete vehicle trajectory is then sent to the onboard terminal of the vehicle traveling on the corresponding road segment, where it is displayed. The complete vehicle trajectory sent to the onboard terminal of the vehicle traveling on the corresponding road segment is the complete vehicle trajectory of vehicles within a preset range, such as 500 meters or 1 kilometer.

[0084] Figure 8 A schematic diagram showing the complete vehicle trajectory displayed on the in-vehicle terminal, such as... Figure 8 As shown, the complete vehicle trajectory of vehicles within a preset range can be displayed in real time on the in-vehicle terminal, allowing the driver to obtain the driving status of vehicles outside their blind spot. It is understandable that... Figure 8 There are no other vehicles around the vehicle where the in-vehicle terminal is located, so only this vehicle is displayed in the in-vehicle terminal.

[0085] Optionally, it can also generate refined cross-sectional flow velocity and density data based on the full-path vehicle trajectory, and provide evidence for customers regarding traffic violations on highways. The full-path vehicle trajectory is the complete vehicle trajectory of all vehicles on all road segments.

[0086] Specifically, the location and speed of vehicles within a road segment can be obtained in real time through real-time vehicle trajectory tracking. This allows for further calculation of traffic flow, speed, and density information for the cross-section or road segment, and also enables the acquisition of license plate attachment results for vehicles violating traffic rules. The license plate attachment results can be obtained by the license plate recognition system installed at the gantry or checkpoint when a vehicle passes through it, and the corresponding vehicle can be identified through the vehicle trajectory.

[0087] Example 2

[0088] Based on the above embodiments, this application embodiment involves a refinement of an implementation method for step S103, which predicts the vehicle trajectory when the vehicle trajectory prediction conditions are met, and fuses the obtained current predicted vehicle trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory. Specifically, it includes the following steps:

[0089] Step S201: When the vehicle trajectory prediction conditions are met, obtain the current vehicle motion trajectory and vehicle driving characteristic information.

[0090] Step S202: Predict the vehicle trajectory based on the current vehicle movement trajectory, vehicle driving characteristic information, and the trained vehicle trajectory prediction model.

[0091] The vehicle driving characteristic information may include first vehicle driving characteristic information extracted from the current vehicle trajectory. This first vehicle driving characteristic information may include the predicted vehicle's position, vehicle type, vehicle speed, and lane information. The predicted vehicle is the vehicle corresponding to the current vehicle trajectory. Vehicle types may include passenger cars, trucks, buses, etc., but this embodiment does not limit the specific types.

[0092] Specifically, vehicle driving feature information is extracted from the current vehicle trajectory, and this vehicle driving feature information is used as input to a trained vehicle trajectory prediction model to obtain the output current vehicle predicted trajectory.

[0093] This application does not limit the method of predicting vehicle trajectories. For example, the vehicle trajectory can be continuously updated based on a trained vehicle trajectory prediction model to predict the vehicle's position until the next sensing device identifies the vehicle's motion trajectory, and then the vehicle trajectory is updated using that motion trajectory. Alternatively, the current vehicle's predicted trajectory within a preset time period can be predicted, and a portion of this predicted trajectory can be fused with the current vehicle's current motion trajectory. For instance, if the current vehicle's predicted trajectory is predicted within 5 seconds, and the next sensing device identifies the vehicle's motion trajectory at the 4th second, then the first 3 seconds of the predicted trajectory within those 5 seconds can be fused with the current vehicle's current motion trajectory.

[0094] In this embodiment, the vehicle trajectory prediction model can be a recurrent neural network (LSTM) or a graph convolutional neural network (GCN). If the vehicle trajectory prediction model is a graph convolutional neural network (GCN), the input to the vehicle trajectory prediction model also includes a road network map. A road network map is a road network topology structure composed of lanes, road segments, intersections, and their upstream and downstream relationships. Therefore, in... Figure 4 It can also preload high-precision road networks.

[0095] It should be understood that before predicting the vehicle trajectory based on the current vehicle trajectory, vehicle driving characteristic information, and the trained vehicle trajectory prediction model, the vehicle trajectory prediction model needs to be trained using the vehicle trajectory as training samples. The method of training the vehicle trajectory prediction model corresponds to the method of predicting the vehicle trajectory based on the trained vehicle trajectory prediction model, and this application embodiment does not specifically limit this.

[0096] The trajectory fusion method provided in this application obtains the current vehicle motion trajectory and vehicle driving feature information when the vehicle trajectory prediction conditions are met; the vehicle trajectory is predicted based on the current vehicle motion trajectory, vehicle driving feature information and a trained vehicle trajectory prediction model to obtain the current vehicle predicted trajectory, and further, the current first fused trajectory can be obtained. Since the prediction of the current vehicle trajectory is relatively accurate when the vehicle trajectory prediction conditions are met, the accuracy of the vehicle trajectory completed when the vehicle is driving in a blind spot position can be improved.

[0097] Optionally, one possible implementation of predicting the vehicle trajectory based on the current vehicle trajectory, vehicle driving characteristic information, and a trained vehicle trajectory prediction model may include the following steps:

[0098] Step S2021: Obtain the first vehicle driving feature information of the vehicle corresponding to the current vehicle motion trajectory at the most recent preset frame number.

[0099] Among them, the first vehicle driving characteristic information includes the vehicle type, position, speed, direction angle, lane information, and other characteristic information of the vehicle corresponding to the current vehicle movement trajectory.

[0100] Step S2022: Obtain the second vehicle driving feature information of the surrounding vehicles in the most recent preset frame.

[0101] In this embodiment, the vehicle driving feature information may further include second vehicle driving feature information of the surrounding vehicles extracted from the vehicle motion trajectories of vehicles surrounding the predicted vehicle. The second vehicle driving feature information may include the vehicle type of the surrounding vehicles, the speed difference, distance difference, and acceleration difference between the surrounding vehicles and the predicted vehicle. The speed difference and distance difference between the surrounding vehicles and the predicted vehicle are vectors.

[0102] Step S2023: Input the first vehicle driving feature information and the second vehicle driving feature information of the most recent preset frame number into the trained vehicle trajectory prediction model.

[0103] Step S2024: Predict the vehicle trajectory using the trained vehicle trajectory prediction model and output the current predicted vehicle trajectory.

[0104] The output of the current vehicle prediction trajectory can be a vehicle prediction trajectory map within a preset time period. The preset time period can be set according to the vehicle trajectory prediction capability, and can be 3 seconds, 5 seconds, 8 seconds, etc., and this embodiment does not specifically limit it.

[0105] Figure 9 This is a schematic diagram illustrating a vehicle trajectory prediction method. The preset frame rate is 10 frames, the preset time is 5 seconds, and the vehicle trajectory prediction model is an LSTM model. Figure 9 X1, X2, X3, ..., X 10 These are the first and second vehicle driving feature information from the most recent 10 frames, respectively. The ellipsis represents X4 to X9, and Y is the predicted trajectory of the current vehicle in 5 seconds. The first and second vehicle driving feature information from the 10 frames can be input into the trained LSTM model to obtain the predicted trajectory of the current vehicle in the next 5 seconds.

[0106] Optionally, based on Embodiment 3, before predicting the vehicle trajectory in step S202 according to the current vehicle trajectory, vehicle driving feature information, and the trained vehicle trajectory prediction model, the method of training the vehicle trajectory prediction model using the vehicle trajectory as training samples can be as follows: (1) Obtain multiple sets of vehicle trajectory training data; the vehicle trajectory training data includes the first vehicle driving feature information of a preset number of frames and the vehicle's real trajectory marked with a preset future time. (2) Iteratively train the vehicle trajectory prediction model according to the multiple sets of vehicle trajectory training data. Specifically, this includes: inputting the training samples into the vehicle trajectory prediction model, adjusting the parameters in the vehicle trajectory prediction model according to the output prediction results and the vehicle's real trajectory, until the error is less than a certain value, and then completing the training of the vehicle trajectory prediction model.

[0107] The trajectory fusion method provided in this application embodiment obtains the first vehicle driving feature information of the vehicle corresponding to the current vehicle motion trajectory within the most recent preset frame number; obtains the second vehicle driving feature information of the surrounding vehicles within the most recent preset frame number; inputs the first and second vehicle driving feature information within the most recent preset frame number into a trained vehicle trajectory prediction model; predicts the vehicle trajectory through the trained vehicle trajectory prediction model and outputs the current vehicle predicted trajectory, thereby realizing the prediction of vehicle trajectory. When the vehicle trajectory prediction conditions are met, the current vehicle predicted trajectory can be obtained through the trained vehicle trajectory prediction model, and a complete vehicle trajectory can be formed using the current vehicle predicted trajectory.

[0108] Example 3

[0109] Based on the above embodiments, this application embodiment involves a refinement of step S104, which involves performing microscopic simulation of the vehicle trajectory when the vehicle trajectory simulation conditions are met, and fusing the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain a current second fused trajectory. Specifically, this includes the following steps:

[0110] Step S301: When the vehicle trajectory simulation conditions are met, obtain the current vehicle trajectory, the current speed of the corresponding vehicle, and the corresponding vehicle type.

[0111] In the implementation of this application, the current vehicle trajectory, the current speed of the corresponding vehicle, and the corresponding vehicle type can be obtained from the sensing device.

[0112] Specifically, image data of the vehicle captured by cameras in the sensing device can be obtained to determine the corresponding vehicle type; the current speed of the corresponding vehicle can be obtained through millimeter-wave radar, lidar, etc. in the sensing device. In addition, the current speed of the vehicle can be calculated based on the vehicle's trajectory.

[0113] Step S302: Determine the current position of the corresponding vehicle based on the current vehicle movement trajectory.

[0114] In this embodiment of the application, the vehicle movement trajectory includes the vehicle's location information, and the current location of the corresponding vehicle can be determined through the current vehicle movement trajectory.

[0115] Step S303: Input the current position, current speed and vehicle type of the corresponding vehicle into the microscopic simulation model.

[0116] Step S304: Perform microscopic simulation of the vehicle trajectory using a microscopic simulation model and output the current vehicle simulation trajectory.

[0117] Among them, the microscopic simulation model is also known as the microscopic traffic simulation model. The microscopic simulation model can process the operation of vehicles on the road network by acquiring information such as the vehicle type, current location, and current speed of vehicles entering the road network, and accurately reflect the interaction between vehicles, such as the interaction during following, lane changes, and driver behavior.

[0118] Specifically, the road network topology and road set conditions are described in advance using a road network description model. The current position, current speed and vehicle type of the corresponding vehicle are input into the micro-simulation model for micro-simulation vehicle departure processing. At least one of the following models, lane changing model, intersection traffic model and parking model in the micro-simulation model is used to perform micro-simulation of the vehicle trajectory and output the current vehicle simulation trajectory.

[0119] The trajectory fusion method provided in this application, when meeting vehicle trajectory simulation conditions, acquires the current vehicle trajectory, the corresponding vehicle's current speed, and the corresponding vehicle type; determines the current position of the corresponding vehicle based on the current vehicle trajectory; inputs the current position, current speed, and vehicle type of the corresponding vehicle into a microscopic simulation model; performs microscopic simulation of the vehicle trajectory using the microscopic simulation model, and outputs the current simulated vehicle trajectory. When a highly accurate predicted trajectory cannot be obtained through vehicle trajectory prediction, the current simulated vehicle trajectory can be obtained through microscopic simulation using the microscopic simulation model. Furthermore, a second fused trajectory can be obtained to complete the vehicle trajectory when it reaches a blind spot, forming a complete vehicle trajectory and avoiding unreasonable phenomena such as collisions and overlaps between the formed complete vehicle trajectories.

[0120] Example 4

[0121] Figure 10 This is a flowchart of a trajectory fusion method provided in Embodiment 4 of this application. Based on the above embodiments, this embodiment involves a refinement of step S404, which uses a microscopic simulation model to perform microscopic simulation of the vehicle trajectory and outputs the current simulated vehicle trajectory. In this embodiment, the microscopic simulation model may include: a car-following model, a lane-changing model, and an intersection traffic model, such as... Figure 10 As shown, the specific steps include:

[0122] Step S401: Using a car-following model, simulate the car-following trajectory of the vehicle based on the current position, current speed and vehicle type of the corresponding vehicle to obtain the current car-following trajectory of the vehicle.

[0123] Specifically, a car-following model is used to follow other vehicles, and the vehicle speed depends on the current speed limit and average speed of the road segment.

[0124] In this embodiment, the following model is first used to simulate the vehicle trajectory and determine whether the corresponding vehicle meets the lane-changing conditions. If the lane-changing conditions are met, step S402 is executed to simulate the lane-changing of the vehicle trajectory using the lane-changing model. When encountering an intersection, it is determined whether the intersection passage conditions are met. If the intersection passage conditions are met, step S403 is executed to simulate the intersection passage using the intersection passage model.

[0125] Step S402: If the corresponding vehicle meets the lane-changing conditions based on the pre-calculated lane-changing probability, then the lane-changing model is used to continue to simulate the lane-changing trajectory of the vehicle along the current vehicle's following trajectory to obtain the current lane-changing simulation trajectory.

[0126] The lane-changing model can simulate the entire process of a vehicle's lane-changing behavior. The content described by the lane-changing model is the entire process of a vehicle's lane-changing behavior, including the generation of the vehicle's lane-changing intention, the feasibility analysis of the lane-changing, and the implementation of the lane-changing behavior.

[0127] In this embodiment, the method for determining whether a vehicle meets the lane-changing condition based on a pre-calculated lane-changing probability is as follows: A random number is used to determine whether the vehicle intends to change lanes. If the vehicle intends to change lanes, the lane-changing condition is met. The random number is determined by the pre-calculated lane-changing probability. For example, the pre-calculated lane-changing probability is 5%. A random number (0 or 1) is generated, with a 5% probability of generating 1. If the generated number is 1, the vehicle intends to change lanes.

[0128] The methods for determining whether a vehicle meets the lane-changing conditions may also include: judging whether the vehicle can enter the adjacent lane. The lane-changing behavior must ensure that there is enough driving space in the adjacent lane. If the insertion of the lane-changing vehicle causes the following vehicle in the adjacent lane to be unable to maintain a normal following state with the inserting vehicle within the minimum reaction time, that is, the headway is less than the emergency following distance, then the lane-changing conditions are not met.

[0129] In this embodiment, if it is determined that the corresponding vehicle meets the lane-changing conditions, a lane-changing model is used to execute the lane change. The lane-changing simulation is performed on the vehicle trajectory along the current vehicle's following trajectory to obtain the current lane-changing simulation trajectory. This embodiment does not limit the method of using a lane-changing model to execute the lane change. For example, a random number can be used to select the lane-changing direction, and then the vehicle trajectory can be simulated according to the selected lane-changing direction.

[0130] Step S403: If the corresponding vehicle meets the intersection passage conditions based on the pre-calculated intersection passage probability, then the intersection passage model is used to continue to simulate the intersection passage of the vehicle trajectory along the current lane-changing simulation trajectory, so as to obtain and output the current vehicle simulation trajectory.

[0131] In this embodiment, the method for determining whether a vehicle meets the intersection passage conditions based on the pre-calculated intersection passage probability is as follows: When determining whether an intersection (i.e., a ramp) has been encountered, if an intersection has been encountered, a random number is used to determine whether the vehicle intends to enter the ramp for merging or diverging. This random number is determined by the pre-calculated intersection passage probability. If the vehicle intends to enter the ramp, the intersection passage conditions are met. For example, the pre-calculated intersection passage probability is 20%. A random number of 0 or 1 is generated, with a 20% probability of generating 1. If the generated number is 1, the vehicle intends to enter the ramp.

[0132] It should be understood that the embodiments of this application do not limit the method of using the intersection traffic model to continue simulating the intersection traffic along the current lane-changing simulation trajectory. For example, the intersection traffic simulation of the vehicle trajectory can be performed based on whether the ramp type is an entrance ramp or an exit ramp, and the distance between the vehicle and the ramp. When simulating the intersection traffic of the vehicle trajectory, path planning and trajectory generation can be performed frame by frame, and the current position of the vehicle can be continuously obtained. If the vehicle's motion trajectory sent by the next sensing device is obtained, it is fused with the vehicle's motion trajectory and the vehicle trajectory is updated using the vehicle's motion trajectory.

[0133] The trajectory fusion method provided in this application employs a car-following model to simulate vehicle trajectories based on the current position, current speed, and vehicle type of the corresponding vehicle, thereby obtaining the current car-following trajectory. If the corresponding vehicle meets the lane-changing conditions based on a pre-calculated lane-changing probability, the lane-changing model continues to simulate lane-changing along the current car-following trajectory, thereby obtaining the current lane-changing simulation trajectory. If the corresponding vehicle meets the intersection passage conditions based on a pre-calculated intersection passage probability, the intersection passage model continues to simulate intersection passage along the current lane-changing simulation trajectory, thereby obtaining and outputting the current vehicle simulation trajectory. Since the pre-calculated intersection passage probability determines whether the corresponding vehicle meets the intersection passage conditions, and the pre-calculated lane-changing probability determines whether the corresponding vehicle meets the lane-changing conditions, the proportion of simulated lane changes or ramp entry by the simulated vehicle can be controlled more accurately, and the traffic flow in blind spots can be simulated more accurately, thus improving the accuracy of micro-simulation of vehicle trajectories to a certain extent.

[0134] Optionally, in this embodiment of the application, before performing microscopic simulation of the vehicle trajectory using the microscopic simulation model and outputting the current vehicle simulation trajectory in step S304, the model parameters in the microscopic simulation model can be calibrated by analyzing the vehicle motion trajectory of the vehicle within the sensing area of ​​the sensing device. This can make the model parameters in the microscopic simulation model more reasonable and improve the accuracy of the vehicle simulation trajectory obtained through the microscopic simulation model.

[0135] For example, based on the above embodiment four, before performing microscopic simulation of the vehicle trajectory using a microscopic simulation model in step S304 and outputting the current vehicle simulation trajectory, it is also necessary to determine the lane-changing probability and the intersection passage probability. In an optional embodiment, the method for determining the lane-changing probability and the intersection passage probability includes the following steps:

[0136] Step S501: Calculate the average probability of lane changing and the average probability of passing through the intersection for all vehicles within the sensing area of ​​the sensing device based on the vehicle movement trajectories of all vehicles.

[0137] Step S502: Determine the average lane-changing probability as the pre-calculated lane-changing probability.

[0138] Step S503: Determine the average intersection passage probability as the pre-calculated intersection passage probability.

[0139] The average probability of lane changing is the average probability of a vehicle changing lanes within the sensing area of ​​the sensing device. The average probability of passing through an intersection is the average probability of entering an intersection to diverge or merge when encountering an intersection within the sensing area of ​​the sensing device.

[0140] An optional method for calculating the average lane-changing probability is as follows: based on the sensing trajectory, obtain the number of all vehicles within the sensing area of ​​the sensing device at a given time point, and determine the number of vehicles that perform lane-changing behavior within the sensing area of ​​the sensing device at that time. The ratio of the number of vehicles that perform lane-changing behavior to the total number of vehicles is determined as the lane-changing probability at that time point. Obtain the lane-changing probabilities at multiple time points, calculate the average of the lane-changing probabilities at multiple time points, and determine the average of the lane-changing probabilities as the average lane-changing probability.

[0141] An optional method for calculating the average probability of intersection passage can be as follows: Based on the sensing trajectory, obtain the number of vehicles that encounter all intersections within the sensing area of ​​the sensing device within a certain time period, and determine the number of vehicles that enter any intersection within the sensing area of ​​the sensing device within that time period. The ratio of the number of vehicles that encounter all intersections within the sensing area of ​​the sensing device within that time period to the number of vehicles that enter any intersection within the sensing area of ​​the sensing device within that time period is determined as the intersection passage probability within that time period. Obtain the intersection passage probabilities within multiple time periods, calculate the average intersection passage probabilities within multiple time periods, and determine the average intersection passage probability as the average probability of intersection passage.

[0142] It should be understood that other methods can also be used to determine the average lane-changing probability and the average intersection passage probability of all vehicles within the sensing area of ​​the computing sensing device, and this application embodiment does not limit this.

[0143] In this embodiment, the average lane-changing probability is determined as a pre-calculated lane-changing probability; the average intersection passage probability is determined as a pre-calculated intersection passage probability. When performing micro-simulation of vehicle trajectories through a micro-simulation model, the corresponding vehicle can be determined to meet the lane-changing condition based on the pre-calculated lane-changing probability, and the corresponding vehicle can be determined to meet the intersection passage condition based on the pre-calculated intersection passage probability. This allows for more accurate control of the proportion of simulated lane changes or entry into ramps by the simulated vehicles, and more accurate simulation of traffic flow in blind spots, thereby improving the accuracy of micro-simulation of vehicle trajectories to a certain extent.

[0144] Example 5

[0145] Based on the above embodiments, this application embodiment relates to a method for determining the current vehicle motion trajectory of the vehicle corresponding to the current vehicle prediction trajectory before step S103 fuses the obtained current vehicle prediction trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory. Specifically, it includes the following steps:

[0146] Step S601: Obtain the first feature of the vehicle corresponding to the current vehicle's predicted trajectory.

[0147] The first feature includes a first attribute feature and a first driving feature. The first attribute feature may include the image features, vehicle type, body color, and body length of the vehicle corresponding to the current vehicle's predicted trajectory. The first driving feature may include the vehicle's position, lane information, and speed information corresponding to the current vehicle's predicted trajectory.

[0148] In this embodiment of the application, the first driving feature of the vehicle corresponding to the current vehicle predicted trajectory can be extracted from the current vehicle predicted trajectory, and the first attribute feature can be extracted from the current vehicle motion trajectory obtained before predicting the vehicle trajectory. The current vehicle motion trajectory may include the vehicle's image information.

[0149] Step S602: Obtain the second feature of the vehicle corresponding to the current vehicle motion trajectory of at least one vehicle.

[0150] The current vehicle trajectory is the vehicle trajectory obtained before the current vehicle prediction trajectory was generated.

[0151] The second feature includes: second attribute features and second driving features. The second attribute features may include the vehicle's image features, vehicle type, body color, and body length corresponding to the current vehicle's trajectory. The second driving features may include the vehicle's position, lane information, and speed information corresponding to the current vehicle's trajectory.

[0152] The image features of the vehicle are those extracted through a deep learning model.

[0153] Step S603: Calculate the vehicle similarity between the first feature and each of the second features.

[0154] This application does not limit the method for calculating the vehicle similarity between the first feature and each of the second features. For example, a weighted summation method can be used to calculate the vehicle similarity between the first feature and each of the second features. A similarity score for each feature can be calculated, such as a vehicle body color similarity score or a vehicle position similarity score, and each feature can be assigned a weight. The vehicle similarity between the first feature and each of the second features is obtained by weighted summation of the feature weights and the similarity scores of each feature.

[0155] Step S604: Determine the current vehicle motion trajectory of the vehicle corresponding to the predicted trajectory of the current vehicle based on the similarity of each vehicle.

[0156] This application does not limit the method for determining the current vehicle trajectory of the vehicle corresponding to the current vehicle's predicted trajectory based on the similarity of each vehicle. For example, a similarity threshold can be set, and a second feature with a vehicle similarity greater than the similarity threshold can be determined. The current vehicle trajectory corresponding to this second feature is then determined as the current vehicle trajectory of the vehicle corresponding to the current vehicle's predicted trajectory. If multiple second features have a vehicle similarity greater than the similarity threshold with the first feature, the second feature with the highest vehicle similarity can be determined, and the current vehicle trajectory corresponding to this second feature is then determined as the current vehicle trajectory of the vehicle corresponding to the current vehicle's predicted trajectory.

[0157] The trajectory fusion method provided in this embodiment can determine the current vehicle motion trajectory of the vehicle corresponding to the current vehicle's predicted trajectory by calculating the vehicle similarity between the first feature and each of the second features, thereby realizing the fusion of the current vehicle's simulated trajectory and the current vehicle motion trajectory of the corresponding vehicle.

[0158] It is understood that before fusing the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle in step S104, the method for determining the current vehicle motion trajectory of the vehicle corresponding to the current vehicle simulation trajectory is similar to the method provided in Embodiment Six of this application, and will not be described in detail here.

[0159] Optionally, based on the above embodiments, after obtaining the current first fused trajectory or the current second fused trajectory, the first fused trajectory or the second fused trajectory can be fused with the next vehicle motion trajectory of the corresponding vehicle, specifically including the following steps:

[0160] Step S701: If the vehicle is detected to move from the blind spot to the sensing area of ​​the sensing device, then the next vehicle movement trajectory is obtained.

[0161] Step S702: Fuse the first fused trajectory or the second fused trajectory with the next vehicle motion trajectory of the corresponding vehicle.

[0162] It is understandable that the method for determining the next vehicle trajectory of the vehicle corresponding to the first fused trajectory or the second fused trajectory can be the same as the method for determining the current vehicle trajectory of the vehicle corresponding to the current vehicle prediction trajectory in Embodiment 5. After determining the next vehicle trajectory of the corresponding vehicle, the first fused trajectory or the second fused trajectory is fused with the next vehicle trajectory of the corresponding vehicle.

[0163] The next vehicle trajectory can be the current vehicle trajectory obtained from any other sensing device. For example, at a first moment, the vehicle trajectory identified by sensing device 1 is acquired, and a predicted trajectory for the current vehicle is formed based on one of the vehicle trajectories; at any time after the first moment, the next vehicle trajectory identified by sensing device 2 for at least one vehicle is acquired. Here, sensing device 2 can be any other sensing device besides sensing device 1.

[0164] The method and steps of fusing the first fused trajectory or the second fused trajectory with the next vehicle motion trajectory of the corresponding vehicle are similar to those of fusing the obtained current vehicle prediction trajectory with the current vehicle motion trajectory of the corresponding vehicle in S103. The embodiments of this application will not be described in detail here.

[0165] The trajectory fusion method provided in this application embodiment, if it detects a vehicle moving from a blind spot to the sensing area of ​​the sensing device, obtains the next vehicle movement trajectory, and fuses the first fused trajectory or the second fused trajectory with the next vehicle movement trajectory of the corresponding vehicle, thereby realizing the fusion of vehicle trajectories from the sensing area to the blind spot and back to the sensing area, completing the vehicle trajectory completion at the blind spot location, effectively solving the problem of vehicle trajectory interruption in scenarios where the sensing device does not have full coverage, and realizing full-path trajectory fusion.

[0166] Example 6

[0167] Figure 11 This is a schematic diagram of the trajectory fusion device provided in Embodiment Six of this application, as shown below. Figure 11 As shown, the trajectory fusion device 80 provided in this embodiment includes: an acquisition module 801, a determination module 802, a prediction module 803, a fusion module 804, and a simulation module 805.

[0168] Specifically, the acquisition module 801 is used to acquire the current vehicle motion trajectory of at least one vehicle identified by the sensing device;

[0169] If module 802 detects a blind zone between the vehicle and the sensing device, it determines that the vehicle trajectory prediction condition or vehicle trajectory simulation condition is met.

[0170] The prediction module 803 predicts the vehicle trajectory when the vehicle trajectory prediction conditions are met.

[0171] The fusion module 804 is used to fuse the obtained current vehicle predicted trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory.

[0172] The simulation module 805 is used to perform microscopic simulation of the vehicle trajectory when the vehicle trajectory simulation conditions are met.

[0173] The fusion module 804 is also used to fuse the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

[0174] The apparatus provided in this application embodiment can be specifically used to execute the method embodiment provided in Embodiment 1 above, and the specific functions will not be repeated here.

[0175] Optionally, the determining module 802 is specifically used to: if a blind zone is detected between the vehicle and the sensing device, determine the blind zone range; if the blind zone range is less than or equal to a first preset blind zone range threshold, determine that the vehicle trajectory prediction condition is met; if the blind zone range is greater than the first preset blind zone range threshold, determine that the vehicle trajectory simulation condition is met.

[0176] Optionally, the prediction module 803 includes an acquisition unit and a prediction unit. The acquisition unit is used to acquire the current vehicle trajectory and vehicle driving characteristic information when the vehicle trajectory prediction conditions are met. The prediction unit is used to predict the vehicle trajectory based on the current vehicle trajectory, vehicle driving characteristic information, and a trained vehicle trajectory prediction model.

[0177] Optionally, the prediction unit is specifically used for: obtaining the first vehicle driving feature information of the vehicle corresponding to the current vehicle's motion trajectory within the most recent preset frame; obtaining the second vehicle driving feature information of the surrounding vehicles within the most recent preset frame; inputting the first and second vehicle driving feature information within the most recent preset frame into a trained vehicle trajectory prediction model; predicting the vehicle trajectory through the trained vehicle trajectory prediction model; and outputting the current vehicle's predicted trajectory.

[0178] Optionally, the simulation module 805 includes a data preparation unit and a simulation unit. The data preparation unit is used to: acquire the current vehicle trajectory, the corresponding vehicle's current speed, and the corresponding vehicle type when the vehicle trajectory simulation conditions are met; determine the current position of the corresponding vehicle based on the current vehicle trajectory; and input the current position, current speed, and vehicle type of the corresponding vehicle into the microscopic simulation model. The simulation unit is used to: perform microscopic simulation of the vehicle trajectory using the microscopic simulation model and output the current vehicle simulation trajectory.

[0179] Optionally, the micro-simulation model includes: a car-following model, a lane-changing model, and an intersection traffic model. The simulation unit is specifically used to: use the car-following model to simulate the vehicle trajectory based on the current position, current speed, and vehicle type of the corresponding vehicle to obtain the current car-following trajectory; if the corresponding vehicle meets the lane-changing conditions based on the pre-calculated lane-changing probability, then use the lane-changing model to continue simulating the lane-changing trajectory along the current car-following trajectory to obtain the current lane-changing simulation trajectory; if the corresponding vehicle meets the intersection traffic conditions based on the pre-calculated intersection traffic probability, then use the intersection traffic model to continue simulating the intersection traffic along the current lane-changing simulation trajectory to obtain and output the current vehicle simulation trajectory.

[0180] Optionally, the trajectory fusion device 80 further includes a probability determination unit. The probability determination unit is used to: acquire the vehicle movement trajectories of all vehicles within the sensing area of ​​the sensing device; calculate the average lane-changing probability and the average intersection passage probability of all vehicles within the sensing area of ​​the sensing device based on the vehicle movement trajectories of all vehicles; determine the average lane-changing probability as a pre-calculated lane-changing probability; and determine the average intersection passage probability as a pre-calculated intersection passage probability.

[0181] Optionally, the determining module 802 is further configured to: obtain a first feature of the vehicle corresponding to the current vehicle predicted trajectory, the first feature including: a first attribute feature and a first driving feature; obtain a second feature of the vehicle corresponding to the current vehicle motion trajectory of at least one vehicle, the second feature including: a second attribute feature and a second driving feature; calculate the vehicle similarity between the first feature and each second feature; and determine the current vehicle motion trajectory of the vehicle corresponding to the current vehicle predicted trajectory based on each vehicle similarity.

[0182] Optionally, the fusion module 804 is further configured to, if it detects that a vehicle is moving from a blind spot into the sensing area of ​​the sensing device, obtain the next vehicle movement trajectory of the vehicle; and fuse the first fused trajectory or the second fused trajectory with the next vehicle movement trajectory of the corresponding vehicle.

[0183] It should be noted that the technical solution and effects of the trajectory fusion device provided in this embodiment can be found in the relevant content of the foregoing method embodiments, and will not be repeated here.

[0184] Example 7

[0185] Figure 12 This is a schematic diagram of the structure of the electronic device provided in Embodiment 7 of this application, as shown below. Figure 12 As shown, this application also provides an electronic device 90, including: a memory 901, a processor 902, and a transceiver 903.

[0186] The memory 901 is used to store programs, and the transceiver 903 is used to send and receive data. Specifically, the program may include program code, which includes computer-executable instructions. The memory 901 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0187] Processor 902 is used to execute programs stored in memory 901.

[0188] The computer program is stored in memory 901 and configured to be executed by processor 902 to implement the trajectory fusion method provided in any embodiment of this application. Related descriptions can be understood by referring to the relevant descriptions and effects corresponding to the steps in the accompanying drawings, and will not be elaborated upon here.

[0189] In this embodiment, the memory 901 and the processor 902 are connected via a bus. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 12 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0190] This application also provides a computer-readable storage medium storing a computer program thereon, which is executed by a processor to implement the trajectory fusion method provided in any embodiment of this application.

[0191] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the trajectory fusion method provided in any embodiment of this application.

[0192] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0193] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0194] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in a combination of hardware and software functional modules.

[0195] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to the processor or controller of a general-purpose computer, special-purpose computer, or other programmable trajectory fusion device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0196] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. 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 of the foregoing.

[0197] Furthermore, although the operations are described in a specific order, this should be understood as requiring that such operations be performed in the specific order shown or in sequential order, or requiring that all illustrated operations be performed to achieve the desired result. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of individual embodiments may also be implemented in combination in a single implementation. Conversely, various features described in the context of a single implementation may also be implemented individually or in any suitable sub-combination in multiple implementations.

[0198] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0199] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A trajectory fusion method, characterized in that, include: Acquire the current vehicle trajectory of at least one vehicle identified by the sensing device; If the vehicle is detected to be traveling into a blind zone between sensing devices, the range of the blind zone is determined; the blind zone is the blind zone between adjacent sensing devices. If the blind zone range is determined to be less than or equal to the first preset blind zone range threshold, then the vehicle trajectory prediction condition is determined to be met. If the blind zone range is determined to be greater than the first preset blind zone range threshold, then the vehicle trajectory simulation conditions are determined to be met. When the vehicle trajectory prediction conditions are met, the vehicle trajectory is predicted, and the obtained current vehicle prediction trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory. When the vehicle trajectory simulation conditions are met, a microscopic simulation of the vehicle trajectory is performed, and the obtained current vehicle simulation trajectory is fused with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

2. The method according to claim 1, characterized in that, The process of predicting the vehicle trajectory when the vehicle trajectory prediction conditions are met includes: When the conditions for vehicle trajectory prediction are met, the current vehicle trajectory and vehicle driving characteristic information are obtained; The vehicle trajectory is predicted based on the current vehicle movement trajectory, vehicle driving characteristic information, and the trained vehicle trajectory prediction model.

3. The method according to claim 2, characterized in that, The process of predicting the vehicle trajectory based on the current vehicle trajectory, vehicle driving characteristic information, and a trained vehicle trajectory prediction model includes: Obtain the first vehicle driving feature information of the vehicle at the most recent preset frame number corresponding to the current vehicle's motion trajectory; Obtain the driving feature information of the second vehicle in the surrounding vehicles within the most recent preset frame number; The first vehicle driving feature information and the second vehicle driving feature information of the most recent preset frame number are input into the trained vehicle trajectory prediction model; The vehicle trajectory is predicted using the trained vehicle trajectory prediction model, and the current predicted vehicle trajectory is output.

4. The method according to claim 1, characterized in that, The process of performing microscopic simulation of the vehicle trajectory when the vehicle trajectory simulation conditions are met includes: When the vehicle trajectory simulation conditions are met, the current vehicle trajectory, the current speed of the corresponding vehicle, and the corresponding vehicle type are obtained. The current position of the corresponding vehicle is determined based on the current vehicle trajectory. Input the current position, current speed, and vehicle type of the corresponding vehicle into the microscopic simulation model; The vehicle trajectory is simulated at a microscopic level using the aforementioned microscopic simulation model, and the current simulated vehicle trajectory is output.

5. The method according to claim 4, characterized in that, The microscopic simulation models include: car-following model, lane-changing model, and intersection traffic model; The step of performing microscopic simulation of the vehicle trajectory using the microscopic simulation model and outputting the current simulated vehicle trajectory includes: The car-following model is used to simulate the car-following trajectory of the vehicle based on the current position, current speed and vehicle type of the corresponding vehicle, so as to obtain the current car-following trajectory of the vehicle. If the corresponding vehicle meets the lane-changing conditions based on the pre-calculated lane-changing probability, then the lane-changing model is used to continue to simulate the lane-changing trajectory of the vehicle along the current vehicle's following trajectory to obtain the current lane-changing simulation trajectory. If the corresponding vehicle meets the intersection passage conditions based on the pre-calculated intersection passage probability, then the intersection passage model is used to continue to simulate the intersection passage of the vehicle trajectory along the current lane-changing simulation trajectory, so as to obtain and output the current vehicle simulation trajectory.

6. The method according to claim 5, characterized in that, Before performing microscopic simulation of the vehicle trajectory using the microscopic simulation model and outputting the current simulated vehicle trajectory, the method further includes: Acquire the vehicle movement trajectories of all vehicles within the sensing area of ​​the sensing device; The average probability of lane changing and the average probability of passing through the intersection for all vehicles within the sensing area of ​​the sensing device are calculated based on the vehicle movement trajectories of all vehicles. The average lane-changing probability is determined to be the pre-calculated lane-changing probability; The average probability of passage at the intersection is determined as the pre-calculated probability of passage at the intersection.

7. The method according to any one of claims 1-6, characterized in that, Before fusing the obtained current vehicle predicted trajectory with the corresponding vehicle's current vehicle motion trajectory to obtain the current first fused trajectory, the method further includes: Obtain the first feature of the vehicle corresponding to the current vehicle's predicted trajectory, the first feature including: a first attribute feature and a first driving feature; Obtain the second feature of the vehicle corresponding to the current vehicle motion trajectory of at least one vehicle, the second feature including: second attribute feature and second driving feature; Calculate the vehicle similarity between the first feature and each of the second features; The current vehicle trajectory is determined based on the similarity of each vehicle.

8. The method according to any one of claims 1-6, characterized in that, Also includes: If the vehicle is detected to have moved from the blind spot into the sensing area of ​​the sensing device, the next vehicle movement trajectory is obtained. The first or second fused trajectory is fused with the next vehicle motion trajectory of the corresponding vehicle.

9. A trajectory fusion device, characterized in that, include: The acquisition module is used to acquire the current vehicle movement trajectory of at least one vehicle identified by the sensing device; The module determines the range of the blind zone if it detects that the vehicle is traveling into the blind zone between the sensing devices. If the blind zone range is determined to be less than or equal to the first preset blind zone range threshold, then the vehicle trajectory prediction condition is determined to be met. If the blind zone range is determined to be greater than the first preset blind zone range threshold, then the vehicle trajectory simulation conditions are determined to be met. The prediction module predicts the vehicle trajectory when the vehicle trajectory prediction conditions are met. The fusion module is used to fuse the obtained current vehicle predicted trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current first fused trajectory; The simulation module is used to perform microscopic simulation of the vehicle trajectory when the vehicle trajectory simulation conditions are met; The fusion module is further configured to fuse the obtained current vehicle simulation trajectory with the current vehicle motion trajectory of the corresponding vehicle to obtain the current second fused trajectory.

10. An electronic device, characterized in that, include: A processor, and a memory and a transceiver communicatively connected to the processor; The memory stores computer-executed instructions; the transceiver is used for sending and receiving data. The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-8.

12. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

Images