Internet-of-vehicles vehicle intelligent cooperative control system based on intelligent driving environment self-adaption

Abstract

The invention discloses an Internet of Vehicles vehicle intelligent cooperative control system based on intelligent driving environment self-adaption, and belongs to the technical field of intelligent driving. The invention discloses an Internet of Vehicles vehicle intelligent cooperative control system based on intelligent driving environment self-adaption. The system comprises an environment perception and data fusion module, an Internet of Vehicles communication module, a self-adaption cooperative planning module and a dynamic control and execution module. The method solves the problem that in the prior art, a driving path is not formulated according to pedestrians and the environment outside a vehicle, and if the pedestrians break into the path suddenly, danger is easily caused, improves the traffic efficiency and safety, reduces the occurrence rate of traffic accidents, reduces the occurrence rate of traffic accidents caused by human errors, and improves the traffic safety. A path planning strategy can be dynamically adjusted by combining future paths of pedestrians and future tracks of vehicles in real time, local optimum of a traditional algorithm can be avoided by introducing technologies such as dynamic exploration factors, the rationality of a global path is improved, and indexes such as time, energy consumption and safety can be optimized at the same time.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and specifically to an intelligent collaborative control system for vehicles in an Internet of Vehicles (IoV) based on adaptive intelligent driving environment. Background Art

[0002] Intelligent driving systems provide driver assistance, including important or useful driving-related information. Typically, these systems are deployed on domain controllers (DCs), accessing data from sensors such as radar and cameras. Through local computation and decision-making, they implement appropriate control of the chassis and powertrain domains.

[0003] The Chinese patent with publication number CN117423227A discloses an intelligent driving collaborative control system based on owner-driver road assistance, including a cloud platform: a roadside terminal: a vehicle terminal: responsible for path planning and control on the vehicle side; specifically: the vehicle-side sensing device perceives the vehicle-side traffic object information; the vehicle intelligent driving control module perceives and integrates the vehicle-side traffic object information to obtain the vehicle-side perception environment model; the vehicle intelligent driving module receives the dispatch instructions issued by the cloud platform; the vehicle intelligent driving module applies to the cloud platform for a static road environment map and an intelligent driving vehicle database of the vehicle area according to the vehicle dispatch instructions, and applies to the edge computing center for a dynamic traffic status map of the area.

[0004] In actual use, the above patent does not formulate a driving path based on pedestrians and the environment outside the vehicle. If a pedestrian suddenly breaks into the path, it is easy to cause danger and has low safety. Therefore, it does not meet existing needs. In response to this, we propose an intelligent collaborative control system for vehicles in the Internet of Vehicles based on intelligent driving environment adaptation. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent collaborative control system for vehicles in the Internet of Vehicles based on intelligent driving environment adaptation. The present invention can obtain and process various vehicle information in real time, realize the collaboration between people, vehicles, roads and clouds, thereby improving traffic efficiency and safety, reducing the incidence of traffic accidents, reducing unnecessary acceleration and braking, and reducing the incidence of traffic accidents caused by human errors. By combining the future paths of pedestrians and the future trajectories of vehicles in real time, the path planning strategy can be dynamically adjusted. By introducing technologies such as dynamic exploration factors, traditional algorithms can be prevented from falling into local optimality, the rationality of the global path can be improved, and indicators such as time, energy consumption, and safety can be optimized at the same time, solving the problems raised in the above background technology.

[0006] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an intelligent collaborative control system for vehicles in the Internet of Vehicles based on intelligent driving environment adaptation, comprising:

[0007] The environmental perception and data fusion module is used to obtain dynamic information about the status of surrounding vehicles and pedestrian movement trajectories, as well as the vehicle's external environment data, and to perform vehicle positioning.

[0008] The Internet of Vehicles communication module is used to realize communication between vehicles, vehicles and infrastructure, vehicles and networks, and vehicles and pedestrians, obtain real-time traffic information, and share vehicle location and speed information;

[0009] The adaptive collaborative planning module is used to combine the vehicle's external environment data, the status of surrounding vehicles, and the movement trajectory of pedestrians to generate multiple dynamic paths through a reinforcement learning algorithm, and then screen the generated multiple dynamic paths to obtain the optimal dynamic path;

[0010] The dynamic control and execution module is used to execute the instructions of the adaptive collaborative planning module to control the steering, acceleration and braking of the vehicle;

[0011] The human-computer interaction and monitoring module is used to provide a human-computer interaction interface, predict potential maintenance needs, perform maintenance in advance, receive real-time traffic information and adjust driving strategies.

[0012] Preferably, the adaptive collaborative planning module specifically includes:

[0013] The model building unit is used to select the neural network architecture, adjust the number of layers and nodes of the neural network, and build the path planning model;

[0014] A prediction unit, which uses the surrounding vehicle states and pedestrian motion trajectories in combination with a neural network to predict the future trajectories of surrounding vehicles and the future paths of pedestrians;

[0015] The path planning unit is used to generate multiple dynamic paths based on the current path planning results combined with the future paths of pedestrians and future trajectories of vehicles, and to screen the generated multiple dynamic paths to obtain the optimal dynamic path.

[0016] Preferably, the path planning process of the adaptive collaborative planning module specifically includes:

[0017] Select the neural network architecture, adjust the number of layers and nodes in the neural network, build a path planning model, and introduce multi-objective optimization items including safety, efficiency, and comfort;

[0018] The surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure after data fusion are combined with vehicle positioning information as state input to output the vehicle path planning result;

[0019] The future trajectory of surrounding vehicles is predicted by combining the state of surrounding vehicles with neural networks, and the future path of pedestrians is predicted by combining the trajectory of pedestrians with neural networks.

[0020] The reinforcement learning algorithm is used to generate multiple dynamic paths based on the current path planning results and the future paths of pedestrians and vehicles. The generated multiple dynamic paths are screened to obtain the optimal dynamic path.

[0021] Preferably, the environment perception and data fusion module includes:

[0022] The collection unit is used to continuously capture dynamic information about the status of surrounding vehicles and pedestrian movement trajectories by deploying millimeter-wave radars at key road nodes, and to collect real-time data about the vehicle's external environment using lidar equipment;

[0023] A fusion unit is used to fuse the vehicle's external environment data, surrounding vehicle status, and pedestrian motion trajectory using a multi-source heterogeneous data fusion algorithm;

[0024] A positioning unit, which is used to determine the vehicle's precise location using GPS, an inertial measurement unit, onboard sensors, and map data.

[0025] Preferably, the collection unit specifically includes:

[0026] Millimeter-wave radars are deployed at key road nodes to continuously emit laser beams to the surrounding environment and calculate the target distance by measuring the time difference of the reflected signals.

[0027] Millimeter-wave radar performs multi-angle detection, generating hundreds of thousands to millions of three-dimensional coordinate points per second to form dense point cloud data.

[0028] Use point cloud data to extract the vehicle's external road boundaries, lane lines, and traffic signs, identify the outlines and distances of pedestrians and roadblocks, and obtain the vehicle's external environment data;

[0029] It uses a multi-beam laser radar, combined with an inertial measurement unit and GPS data, and uses Kalman filtering to eliminate motion blur, track vehicle information in real time, and obtain the status of surrounding vehicles;

[0030] Track the displacement and speed changes of dynamic pedestrian targets in real time and use semantic segmentation models to identify pedestrian motion trajectories;

[0031] The edge computing nodes perform preliminary screening of the vehicle's external environment data, surrounding vehicle status, and pedestrian movement trajectories to eliminate signal noise caused by weather interference or equipment aging.

[0032] The collected data on the vehicle's external environment, surrounding vehicle status, and pedestrian movement trajectory are denoised and missing values ​​are interpolated. The wavelet threshold method is used to process the time series data noise, the timestamps of different devices are aligned, and the Kalman filter is applied to correct the trajectory drift error.

[0033] Preferably, the fusion unit comprises:

[0034] A spatiotemporal alignment algorithm is used to fuse the collected vehicle external environment data, surrounding vehicle status, and pedestrian motion trajectory, and a sliding window mechanism is used to eliminate instantaneous fluctuation errors.

[0035] Identify and correct abnormal trajectory points and convert traffic flow data of different formats into a standardized space-time matrix;

[0036] Convert image data into tensor structure, parse trajectory data into GeoJson format, and encode pedestrian motion trajectory into spatiotemporal matrix;

[0037] Extract multi-scale image features, combine the PointPillars network to process point clouds to generate pseudo-image feature maps, and analyze the frequency domain characteristics of vehicle dynamics data through power spectral density;

[0038] Based on the spatiotemporal clustering of pedestrian detection frames, the pedestrian residence time, aggregation density and movement direction distribution characteristics are extracted. The extracted features are weighted fused to obtain the fused surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure data.

[0039] Preferably, the dynamic control and execution module includes:

[0040] The execution unit is used to convert the planned optimal dynamic path into specific steering, acceleration and braking control instructions;

[0041] The control module is used to control the steering, acceleration and braking of the vehicle according to the control instructions of the execution unit.

[0042] Preferably, the human-computer interaction and monitoring module includes:

[0043] User interface unit, used to provide a human-computer interaction interface, display path planning results, vehicle driving status and warning information, and provide navigation services;

[0044] Safety response unit, which is used to monitor and analyze vehicle operating data, predict potential maintenance needs based on the analysis results, and ensure driving safety;

[0045] Traffic management unit, used to communicate with the traffic management center, receive real-time traffic information and adjust driving strategies.

[0046] Preferably, after positioning the vehicle, the method further comprises:

[0047] Collect multi-angle images of the vehicle from the first-person perspective and determine the vehicle's current visual position based on the multi-angle images;

[0048] Determine the positioning perspective pose of the vehicle in the third-person perspective based on the vehicle positioning result, and determine the posture deviation between the current perspective pose and the positioning perspective pose;

[0049] Determine the response state deviation of the vehicle positioning accuracy at multiple angles based on the posture deviation, and obtain the vehicle's offset data in the front, back, left, and right directions based on the response state deviation;

[0050] Constructing a vehicle positioning deviation spatial distribution table based on the offset data, and determining the position deviation level of the vehicle in each direction of front, back, left, and right based on the positioning deviation spatial distribution table;

[0051] Determine the vehicle's posture deviation matrix in each direction of the front, back, left, and right based on the position deviation level, and determine the vehicle's deviation vector in each direction of the front, back, left, and right based on the posture deviation matrix;

[0052] Determine a spatial impulse response deviation parameter of the vehicle based on the deviation vector, and determine an information perception delay error of a driving influencing factor of the vehicle according to the spatial impulse response deviation parameter;

[0053] Construct an observation delay model for vehicle positioning perception based on information perception delay error;

[0054] The actual positioning parameters of the vehicle are output according to the positioning results of the vehicle through the observation delay model, and the actual positioning parameters are used as the final positioning results of the vehicle.

[0055] Preferably, the generated multiple dynamic paths are screened to obtain the optimal dynamic path, specifically including:

[0056] Obtain the number of intersections on each dynamic path and the vehicle flow parameters and pedestrian flow parameters of each intersection according to the electronic map;

[0057] Determine the traffic capacity value of each intersection based on vehicle flow parameters and pedestrian flow parameters;

[0058] Obtain the distribution of high-density vehicle flow groups on each dynamic path and the congestion scale of high-density vehicle flow groups through electronic navigation equipment;

[0059] Obtain the blocking density of the high-density vehicle flow group according to the blocking scale of the high-density vehicle flow group;

[0060] The recommendation index of each dynamic path is calculated based on the traffic capacity value of each intersection and the blocking density of high-density vehicle flow groups:

[0061]

[0062] Among them, F jis the recommendation index of the j-th dynamic path, θ is the traffic efficiency weight, which is 0.6, P is the traffic capacity decision factor, M is the number of high-density vehicle flow groups, N is the number of intersections, Nj is the number of intersections on the j-th dynamic path, i is the i-th intersection, σ i is the traffic capacity value of the i-th intersection, Mj is the number of high-density vehicle flow groups on the j-th dynamic path, k is the k-th high-density vehicle flow group, μ k It is represented as the blocking density of the kth high-density vehicle flow group, δ is represented as the traffic travel time weight, which is 0.4, l j Expressed as the path length of the jth dynamic path, v j It is expressed as the fastest speed of the vehicle on the jth dynamic path, s j It is expressed as the safe braking distance when the vehicle is traveling at the fastest speed on the jth dynamic path, d j Expressed as the average distance between vehicles on the jth dynamic path, t j It is represented as the standard driving time on the jth dynamic path;

[0063] The dynamic path with the largest recommendation index is selected as the optimal dynamic path.

[0064] Compared with the prior art, the present invention has the following beneficial effects:

[0065] Through vehicle networking technology, the present invention can obtain and process various vehicle information in real time, realize human-vehicle-road-cloud collaboration, thereby improving traffic efficiency and safety, reducing the incidence of traffic accidents, and can optimize driving routes according to real-time road conditions information, reduce unnecessary acceleration and braking, and use millimeter-wave radar to perceive the surrounding environment in all directions, formulate the optimal dynamic path, and reduce the incidence of traffic accidents caused by human error. Through precise perception and intelligent decision-making, the vehicle can adjust the driving strategy in real time to respond to emergencies. By combining the future paths of pedestrians and the future trajectories of vehicles in real time, it can dynamically adjust the path planning strategy to adapt to complex and changing traffic environments. The reinforcement learning framework can quickly balance immediate benefits with long-term goals to generate efficient paths. The introduction of technologies such as dynamic exploration factors can prevent traditional algorithms from falling into local optimality, improve the rationality of the global path, and can simultaneously optimize indicators such as time, energy consumption, and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a schematic diagram of the vehicle intelligent collaborative control system module based on intelligent driving environment adaptation of the Internet of Vehicles of the present invention;

[0067] Figure 2 This is a flow chart of the intelligent collaborative control system of the Internet of Vehicles based on intelligent driving environment adaptation of the present invention. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0069] To solve the problem that existing technologies do not formulate driving paths based on pedestrians and the environment outside the vehicle, and if pedestrians suddenly break into the path, it is easy to cause danger, please refer to Figure 1-Figure 2 , this embodiment provides the following technical solutions:

[0070] The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation includes:

[0071] The environmental perception and data fusion module is used to obtain dynamic information about the status of surrounding vehicles and pedestrian movement trajectories, as well as the vehicle's external environment data, and to perform vehicle positioning.

[0072] The IoV communication module is used to enable communication between vehicles, vehicles and infrastructure, vehicles and networks, and vehicles and pedestrians, obtain real-time traffic information, and share vehicle location and speed information to improve driving safety and efficiency;

[0073] The adaptive collaborative planning module combines the vehicle's external environment data, the status of surrounding vehicles, and the movement trajectories of pedestrians to generate multiple dynamic paths through a reinforcement learning algorithm. The module then screens the generated dynamic paths to obtain the optimal dynamic path, achieving global optimal decisions for lane changing, following vehicles, and obstacle avoidance.

[0074] The dynamic control and execution module is used to execute the instructions of the adaptive collaborative planning module to control the vehicle's steering, acceleration, and braking to achieve autonomous driving;

[0075] The human-computer interaction and monitoring module is used to provide a human-computer interaction interface, predict potential maintenance needs, perform maintenance in advance, receive real-time traffic information and adjust driving strategies.

[0076] Through vehicle networking technology, various vehicle information can be obtained and processed in real time, and human-vehicle-road-cloud collaboration can be achieved, thereby improving traffic efficiency and safety. Vehicles communicate and exchange data with other vehicles, infrastructure and pedestrians, which can optimize traffic flow, reduce traffic congestion, and reduce the incidence of traffic accidents. Drivers can obtain road conditions information, navigation services and entertainment content in real time, making the driving process more convenient and comfortable. In addition, intelligent connected vehicles can also provide multi-vehicle collaborative control, improve road utilization, and further optimize the driving experience. Intelligent connected vehicles can optimize driving routes based on real-time road conditions information, reduce unnecessary acceleration and braking, thereby reducing energy consumption and emissions, and help achieve green travel. Millimeter-wave radar can fully perceive the surrounding environment and reduce the incidence of traffic accidents caused by human error. Through precise perception and intelligent decision-making, vehicles can adjust driving strategies in real time to respond to emergencies. In highway scenarios, vehicle-road collaboration optimizes vehicle driving safety and traffic efficiency; at urban intersections, it improves pedestrian protection and reduces collision risks.

[0077] Adaptive collaborative planning module, specifically including:

[0078] The model building unit is used to select the neural network architecture, adjust the number of layers and nodes of the neural network, and build the path planning model;

[0079] A prediction unit, which uses the surrounding vehicle states and pedestrian motion trajectories in combination with a neural network to predict the future trajectories of surrounding vehicles and the future paths of pedestrians;

[0080] The path planning unit is used to generate multiple dynamic paths based on the current path planning results and the future paths of pedestrians and vehicles, and to screen the generated multiple dynamic paths to obtain the optimal dynamic path;

[0081] The path planning process of the adaptive collaborative planning module includes:

[0082] Select the neural network architecture, adjust the number of layers and nodes in the neural network, build a path planning model, and introduce multi-objective optimization items including safety, efficiency, and comfort;

[0083] The surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure after data fusion are combined with vehicle positioning information as state input to output the vehicle path planning result;

[0084] The future trajectory of surrounding vehicles is predicted by combining the state of surrounding vehicles with neural networks, and the future path of pedestrians is predicted by combining the trajectory of pedestrians with neural networks.

[0085] The reinforcement learning algorithm is used to generate multiple dynamic paths based on the current path planning results and the future paths of pedestrians and vehicles. The generated multiple dynamic paths are screened to obtain the optimal dynamic path.

[0086] By combining the future paths of pedestrians and vehicles in real time, the path planning strategy can be dynamically adjusted to adapt to complex and changing traffic environments. The reinforcement learning framework can quickly balance immediate benefits with long-term goals to generate efficient paths. The introduction of technologies such as dynamic exploration factors can prevent traditional algorithms from falling into local optimality and improve the rationality of the global path. It can simultaneously optimize indicators such as time, energy consumption, and safety. For example, through the design of reward functions, it can balance benefits and risks. The reinforcement learning algorithm is robust to random events (such as the sudden appearance of pedestrians) and is suitable for trajectory generation in dynamic environments.

[0087] Environmental perception and data fusion module, including:

[0088] The collection unit is used to continuously capture dynamic information about the status of surrounding vehicles and pedestrian movement trajectories by deploying millimeter-wave radars at key road nodes, and to collect real-time data about the vehicle's external environment using lidar equipment;

[0089] A fusion unit is used to fuse the vehicle's external environment data, surrounding vehicle status, and pedestrian motion trajectory using a multi-source heterogeneous data fusion algorithm;

[0090] The positioning unit is used to determine the vehicle's precise location through GPS, inertial measurement unit, on-board sensors and map data, ensuring accurate navigation of the vehicle in complex environments.

[0091] The collection unit specifically includes:

[0092] Millimeter-wave radars are deployed at key road nodes to continuously emit laser beams to the surrounding environment and calculate the target distance by measuring the time difference of the reflected signals.

[0093] Millimeter-wave radar performs multi-angle detection, generating hundreds of thousands to millions of three-dimensional coordinate points per second to form dense point cloud data.

[0094] Use point cloud data to extract the vehicle's external road boundaries, lane lines, and traffic signs, identify the outlines and distances of pedestrians and roadblocks, and obtain the vehicle's external environment data;

[0095] It uses a multi-beam laser radar, combined with an inertial measurement unit and GPS data, and uses Kalman filtering to eliminate motion blur, track vehicle information in real time, and obtain the status of surrounding vehicles;

[0096] Track the displacement and speed changes of dynamic pedestrian targets in real time and use semantic segmentation models to identify pedestrian motion trajectories;

[0097] The edge computing nodes perform preliminary screening of the vehicle's external environment data, surrounding vehicle status, and pedestrian movement trajectories to eliminate signal noise caused by weather interference or equipment aging.

[0098] The collected data on the vehicle's external environment, surrounding vehicle status, and pedestrian movement trajectory are denoised and missing values ​​are interpolated. The wavelet threshold method is used to process the time series data noise, the timestamps of different devices are aligned, and the Kalman filter is applied to correct the trajectory drift error.

[0099] Fusion unit, including:

[0100] A spatiotemporal alignment algorithm is used to fuse the collected vehicle external environment data, surrounding vehicle status, and pedestrian motion trajectory, and a sliding window mechanism is used to eliminate instantaneous fluctuation errors.

[0101] Identify and correct abnormal trajectory points and convert traffic flow data of different formats into a standardized space-time matrix;

[0102] Convert image data into tensor structure, parse trajectory data into GeoJson format, and encode pedestrian motion trajectory into spatiotemporal matrix;

[0103] Extract multi-scale image features, combine the PointPillars network to process point clouds to generate pseudo-image feature maps, and analyze the frequency domain characteristics of vehicle dynamics data through power spectral density;

[0104] Based on the spatiotemporal clustering of pedestrian detection frames, the pedestrian residence time, aggregation density and movement direction distribution characteristics are extracted. The extracted features are weighted fused to obtain the fused surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure data.

[0105] Dynamic control and execution module, including:

[0106] The execution unit is used to convert the planned optimal dynamic path into specific steering, acceleration, and braking control instructions to ensure that the vehicle follows the optimal dynamic path;

[0107] The control module is used to control the vehicle's steering, acceleration, and braking according to the control instructions of the execution unit to achieve automatic driving. It adjusts the control parameters through a closed-loop feedback mechanism to ensure driving smoothness and safety and achieve synchronous coordination of vehicle spacing and speed.

[0108] Human-computer interaction and monitoring module, including:

[0109] User interface unit, used to provide a human-computer interaction interface, display path planning results, vehicle driving status and warning information, provide navigation services, and ensure that drivers or passengers can understand the vehicle status in real time;

[0110] Safety response unit, which monitors and analyzes vehicle operating data, predicts potential maintenance needs based on the analysis results, ensures driving safety, and reduces the occurrence of failures;

[0111] The traffic management unit is used to communicate with the traffic management center, receive real-time traffic information and adjust driving strategies to improve driving efficiency.

[0112] Working principle: When using the intelligent collaborative control system of the vehicle network based on intelligent driving environment adaptation of the present invention, according to Figure 1 and Figure 2 , including the following steps:

[0113] S1: Obtain dynamic information about surrounding vehicle status, pedestrian movement trajectories, and the vehicle's external environment data for integration, while simultaneously positioning the vehicle and transmitting the data to the adaptive collaborative planning module using the Internet of Vehicles communication module;

[0114] S2: The adaptive collaborative planning module selects the neural network architecture, adjusts the number of layers and nodes in the neural network, builds a path planning model, and introduces multi-objective optimization items including safety, efficiency, and comfort.

[0115] S3: Combines the fused data of surrounding vehicle trajectories, pedestrian movement trajectories, and road topology with vehicle positioning information as state input, outputs the vehicle's path planning results, and uses neural networks to predict the future trajectories of surrounding vehicles and pedestrian paths.

[0116] S4: Using a reinforcement learning algorithm to generate multiple dynamic paths based on the current path planning results and the future paths of pedestrians and vehicles, the generated multiple dynamic paths are screened to obtain the optimal dynamic path;

[0117] S5: Convert the planned optimal dynamic path into specific steering, acceleration, and braking control instructions, and control the steering, acceleration, and braking of the vehicle according to the control instructions of the execution unit.

[0118] In summary, the intelligent collaborative control system of the vehicle network based on intelligent driving environment adaptation of the present invention can obtain and process various vehicle information in real time through vehicle network technology, realize the collaboration of people, vehicles, roads and clouds, thereby improving traffic efficiency and safety. Vehicles communicate and exchange data with other vehicles, infrastructure and pedestrians, which can optimize traffic flow, reduce traffic congestion, and reduce the incidence of traffic accidents. Drivers can obtain road condition information, navigation services and entertainment content in real time, making the driving process more convenient and comfortable. In addition, intelligent networked vehicles can also provide multi-vehicle collaborative control, improve road utilization, and further optimize the driving experience. Intelligent networked vehicles can optimize driving routes according to real-time road condition information, reduce unnecessary acceleration and braking, thereby reducing energy consumption and emissions, and helping To achieve green travel, millimeter-wave radar can fully perceive the surrounding environment and reduce the incidence of traffic accidents caused by human errors. Through precise perception and intelligent decision-making, vehicles can adjust driving strategies in real time to respond to emergencies. In highway scenarios, vehicle-road collaboration optimizes vehicle driving safety and traffic efficiency. At urban intersections, it improves pedestrian protection and reduces collision risks. By combining the future paths of pedestrians and the future trajectories of vehicles in real time, it can dynamically adjust the path planning strategy to adapt to complex and changing traffic environments. The reinforcement learning framework can quickly balance immediate benefits with long-term goals and generate efficient paths. The introduction of technologies such as dynamic exploration factors can prevent traditional algorithms from falling into local optimality, improve the rationality of the global path, and simultaneously optimize time, energy consumption, safety and other indicators.

[0119] In one embodiment, after positioning the vehicle, the method further includes:

[0120] Collect multi-angle images of the vehicle from the first-person perspective and determine the vehicle's current visual position based on the multi-angle images;

[0121] Determine the positioning perspective pose of the vehicle in the third-person perspective based on the vehicle positioning result, and determine the posture deviation between the current perspective pose and the positioning perspective pose;

[0122] Determine the response state deviation of the vehicle positioning accuracy at multiple angles based on the posture deviation, and obtain the vehicle's offset data in the front, back, left, and right directions based on the response state deviation;

[0123] Constructing a vehicle positioning deviation spatial distribution table based on the offset data, and determining the position deviation level of the vehicle in each direction of front, back, left, and right based on the positioning deviation spatial distribution table;

[0124] Determine the vehicle's posture deviation matrix in each direction of the front, back, left, and right based on the position deviation level, and determine the vehicle's deviation vector in each direction of the front, back, left, and right based on the posture deviation matrix;

[0125] Determine a spatial impulse response deviation parameter of the vehicle based on the deviation vector, and determine an information perception delay error of a driving influencing factor of the vehicle according to the spatial impulse response deviation parameter;

[0126] Construct an observation delay model for vehicle positioning perception based on information perception delay error;

[0127] The actual positioning parameters of the vehicle are output according to the positioning results of the vehicle through the observation delay model, and the actual positioning parameters are used as the final positioning results of the vehicle.

[0128] In this embodiment, the response state deviation is represented by the position state response deviation of the vehicle at each angle;

[0129] In this embodiment, the positioning deviation spatial distribution table is represented as a distribution table of various angles of the vehicle's positioning deviation in the road space;

[0130] In this embodiment, the spatial impact response deviation parameter is represented by an impact time delay response deviation parameter when a vehicle collides with another vehicle or object in a road space.

[0131] The beneficial effects of the above technical solution are: by determining the positioning deviation of the vehicle from different perspectives to construct an observation delay model to correct and adjust the positioning results, the high accuracy and reliability of the vehicle positioning results can be guaranteed, and accurate reference data can be provided for subsequent path planning and determination of relative positions with other vehicles and people, thereby improving practicality and reliability.

[0132] In one embodiment, the generated multiple dynamic paths are screened to obtain the optimal dynamic path, specifically including:

[0133] Obtain the number of intersections on each dynamic path and the vehicle flow parameters and pedestrian flow parameters of each intersection according to the electronic map;

[0134] Determine the traffic capacity value of each intersection based on vehicle flow parameters and pedestrian flow parameters;

[0135] Obtain the distribution of high-density vehicle flow groups on each dynamic path and the congestion scale of high-density vehicle flow groups through electronic navigation equipment;

[0136] Obtain the blocking density of the high-density vehicle flow group according to the blocking scale of the high-density vehicle flow group;

[0137] The recommendation index of each dynamic path is calculated based on the traffic capacity value of each intersection and the blocking density of high-density vehicle flow groups:

[0138]

[0139] Among them, F jis the recommendation index of the j-th dynamic path, θ is the traffic efficiency weight, which is 0.6, P is the traffic capacity decision factor, M is the number of high-density vehicle flow groups, N is the number of intersections, Nj is the number of intersections on the j-th dynamic path, i is the i-th intersection, σ i is the traffic capacity value of the i-th intersection, Mj is the number of high-density vehicle flow groups on the j-th dynamic path, k is the k-th high-density vehicle flow group, μ k It is represented as the blocking density of the kth high-density vehicle flow group, δ is represented as the traffic travel time weight, which is 0.4, l j Expressed as the path length of the jth dynamic path, v j It is expressed as the fastest speed of the vehicle on the jth dynamic path, s j It is expressed as the safe braking distance when the vehicle is traveling at the fastest speed on the jth dynamic path, d j Expressed as the average distance between vehicles on the jth dynamic path, t j It is represented as the standard driving time on the jth dynamic path;

[0140] The dynamic path with the largest recommendation index is selected as the optimal dynamic path.

[0141] In this embodiment, the recommendation index is expressed as the comprehensive recommendation degree of each dynamic path.

[0142] In this embodiment, the traffic capacity decision factor is represented by the influencing factors considered when evaluating road capacity, such as vehicle speed distribution and lane width.

[0143] In this embodiment, the traffic capacity value is expressed as the maximum number of vehicles that can pass through a certain road section or intersection within a unit time.

[0144] In this embodiment, a high-density vehicle flow group represents a traffic state in which the number of vehicles per unit length is relatively large.

[0145] In this embodiment, the congestion density of a high-density vehicle flow group is represented by the congestion degree of a road section with a large number of vehicles.

[0146] The beneficial effects of the above technical solution are: by calculating the recommendation index of each dynamic path according to the traffic efficiency weight and the traffic time weight, the priority of each dynamic path can be intuitively evaluated based on the traffic status of each dynamic path, which can improve the travel efficiency of vehicles and reduce congestion. At the same time, it can enhance the user's travel experience and ensure user satisfaction.

[0147] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0148] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation is characterized by: include: The environmental perception and data fusion module is used to obtain dynamic information about the status of surrounding vehicles and pedestrian movement trajectories, as well as the vehicle's external environment data, and to perform vehicle positioning. The Internet of Vehicles communication module is used to realize communication between vehicles, vehicles and infrastructure, vehicles and networks, and vehicles and pedestrians, obtain real-time traffic information, and share vehicle location and speed information; The adaptive collaborative planning module is used to combine the vehicle's external environment data, the status of surrounding vehicles, and the movement trajectory of pedestrians to generate multiple dynamic paths through a reinforcement learning algorithm, and then screen the generated multiple dynamic paths to obtain the optimal dynamic path; The dynamic control and execution module is used to execute the instructions of the adaptive collaborative planning module to control the steering, acceleration and braking of the vehicle; The human-computer interaction and monitoring module is used to provide a human-computer interaction interface, predict potential maintenance needs, perform maintenance in advance, receive real-time traffic information and adjust driving strategies.

2. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1 is characterized in that: After locating the vehicle, it also includes: Collect multi-angle images of the vehicle from the first-person perspective and determine the vehicle's current visual position based on the multi-angle images; Determine the positioning perspective pose of the vehicle in the third-person perspective based on the vehicle positioning result, and determine the posture deviation between the current perspective pose and the positioning perspective pose; Determine the response state deviation of the vehicle positioning accuracy at multiple angles based on the posture deviation, and obtain the vehicle's offset data in the front, back, left, and right directions based on the response state deviation; Constructing a vehicle positioning deviation spatial distribution table based on the offset data, and determining the position deviation level of the vehicle in each direction of front, back, left, and right based on the positioning deviation spatial distribution table; Determine the vehicle's posture deviation matrix in each direction of the front, back, left, and right based on the position deviation level, and determine the vehicle's deviation vector in each direction of the front, back, left, and right based on the posture deviation matrix; Determine a spatial impulse response deviation parameter of the vehicle based on the deviation vector, and determine an information perception delay error of a driving influencing factor of the vehicle according to the spatial impulse response deviation parameter; Construct an observation delay model for vehicle positioning perception based on information perception delay error; The actual positioning parameters of the vehicle are output according to the positioning results of the vehicle through the observation delay model, and the actual positioning parameters are used as the final positioning results of the vehicle.

3. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1, characterized in that: The screening of the generated multiple dynamic paths to obtain the optimal dynamic path specifically includes: Obtain the number of intersections on each dynamic path and the vehicle flow parameters and pedestrian flow parameters of each intersection according to the electronic map; Determine the traffic capacity value of each intersection based on vehicle flow parameters and pedestrian flow parameters; Obtain the distribution of high-density vehicle flow groups on each dynamic path and the congestion scale of high-density vehicle flow groups through electronic navigation equipment; Obtain the blocking density of the high-density vehicle flow group according to the blocking scale of the high-density vehicle flow group; The recommendation index of each dynamic path is calculated based on the traffic capacity value of each intersection and the blocking density of high-density vehicle flow groups: Among them, F j is the recommendation index of the j-th dynamic path, θ is the traffic efficiency weight, which is 0.6, P is the traffic capacity decision factor, M is the number of high-density vehicle flow groups, N is the number of intersections, Nj is the number of intersections on the j-th dynamic path, i is the i-th intersection, σ i is the traffic capacity value of the i-th intersection, Mj is the number of high-density vehicle flow groups on the j-th dynamic path, k is the k-th high-density vehicle flow group, μ k It is represented as the blocking density of the kth high-density vehicle flow group, δ is represented as the traffic travel time weight, which is 0.4, l j Expressed as the path length of the jth dynamic path, v j It is expressed as the fastest speed of the vehicle on the jth dynamic path, s j It is expressed as the safe braking distance when the vehicle is traveling at the fastest speed on the jth dynamic path, d j Expressed as the average distance between vehicles on the jth dynamic path, t j It is represented as the standard driving time on the jth dynamic path; The dynamic path with the largest recommendation index is selected as the optimal dynamic path.

4. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1, characterized in that: The adaptive collaborative planning module specifically includes: The model building unit is used to select the neural network architecture, adjust the number of layers and nodes of the neural network, and build the path planning model; A prediction unit, which uses the surrounding vehicle states and pedestrian motion trajectories in combination with a neural network to predict the future trajectories of surrounding vehicles and the future paths of pedestrians; The path planning unit is used to generate multiple dynamic paths based on the current path planning results combined with the future paths of pedestrians and future trajectories of vehicles, and to screen the generated multiple dynamic paths to obtain the optimal dynamic path.

5. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1, characterized in that: The path planning process of the adaptive collaborative planning module specifically includes: Select the neural network architecture, adjust the number of layers and nodes in the neural network, build a path planning model, and introduce multi-objective optimization items including safety, efficiency, and comfort; The surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure after data fusion are combined with vehicle positioning information as state input to output the vehicle path planning result; The future trajectory of surrounding vehicles is predicted by combining the state of surrounding vehicles with neural networks, and the future path of pedestrians is predicted by combining the trajectory of pedestrians with neural networks. The reinforcement learning algorithm is used to generate multiple dynamic paths based on the current path planning results and the future paths of pedestrians and vehicles. The generated multiple dynamic paths are screened to obtain the optimal dynamic path.

6. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1, characterized in that: The environment perception and data fusion module includes: The collection unit is used to continuously capture dynamic information about the status of surrounding vehicles and pedestrian movement trajectories by deploying millimeter-wave radars at key road nodes, and to collect real-time data about the vehicle's external environment using lidar equipment; A fusion unit is used to fuse the vehicle's external environment data, surrounding vehicle status, and pedestrian motion trajectory using a multi-source heterogeneous data fusion algorithm; A positioning unit is used to determine the vehicle's precise location using GPS, an inertial measurement unit, onboard sensors, and map data.

7. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 6 is characterized in that: The collecting unit specifically includes: Millimeter-wave radars are deployed at key road nodes to continuously emit laser beams to the surrounding environment and calculate the target distance by measuring the time difference of the reflected signals. Millimeter-wave radar performs multi-angle detection, generating hundreds of thousands to millions of three-dimensional coordinate points per second to form dense point cloud data. Use point cloud data to extract the vehicle's external road boundaries, lane lines, and traffic signs, identify the outlines and distances of pedestrians and roadblocks, and obtain the vehicle's external environment data; It uses a multi-beam laser radar, combined with an inertial measurement unit and GPS data, and uses Kalman filtering to eliminate motion blur, track vehicle information in real time, and obtain the status of surrounding vehicles; Track the displacement and speed changes of dynamic pedestrian targets in real time and use semantic segmentation models to identify pedestrian movement trajectories; The edge computing nodes perform preliminary screening of the vehicle's external environment data, surrounding vehicle status, and pedestrian movement trajectories to eliminate signal noise caused by weather interference or equipment aging. The collected data on the vehicle's external environment, surrounding vehicle status, and pedestrian movement trajectory are denoised and missing values ​​are interpolated. The wavelet threshold method is used to process the time series data noise, the timestamps of different devices are aligned, and the Kalman filter is applied to correct the trajectory drift error.

8. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 6 is characterized in that: The fusion unit comprises: A spatiotemporal alignment algorithm is used to fuse the collected vehicle external environment data, surrounding vehicle status, and pedestrian movement trajectory, and a sliding window mechanism is used to eliminate instantaneous fluctuation errors. Identify and correct abnormal trajectory points and convert traffic flow data of different formats into a standardized space-time matrix; Convert image data into tensor structure, parse trajectory data into GeoJson format, and encode pedestrian motion trajectory into spatiotemporal matrix; Extract multi-scale image features, combine the PointPillars network to process point clouds to generate pseudo-image feature maps, and analyze the frequency domain characteristics of vehicle dynamics data through power spectral density; Based on the spatiotemporal clustering of pedestrian detection frames, the pedestrian residence time, aggregation density and movement direction distribution characteristics are extracted. The extracted features are weighted fused to obtain the fused surrounding vehicle trajectories, pedestrian movement trajectories and road topology structure data.

9. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 1, characterized in that: The dynamic control and execution module includes: The execution unit is used to convert the planned optimal dynamic path into specific steering, acceleration and braking control instructions; The control module is used to control the steering, acceleration and braking of the vehicle according to the control instructions of the execution unit.

10. The intelligent collaborative control system for connected vehicles based on intelligent driving environment adaptation according to claim 6, characterized in that: The human-computer interaction and monitoring module includes: User interface unit, used to provide a human-computer interaction interface, display path planning results, vehicle driving status and warning information, and provide navigation services; Safety response unit, which is used to monitor and analyze vehicle operating data and predict potential maintenance needs based on the analysis results; Traffic management unit, used to communicate with the traffic management center, receive real-time traffic information and adjust driving strategies.

Images