Automatic driving system and method integrating vehicle-road cooperation and social network

Abstract

The invention discloses an automatic driving system integrating vehicle-road cooperation and a social network, and the system comprises a data acquisition module which is used for collecting the information data of the environment around a vehicle, and the information data comprises the sensor data of the vehicle and an external data source; the communication module is used for transmitting the acquired environment information data through a honeycomb car networking technology; the decision module is used for performing comprehensive analysis and decision making on the acquired environment information data through a deep learning model, accurately predicting the current road condition and generating a speed control scheme; and the execution module is used for controlling the actual movement of the vehicle according to the specified decision output instruction. The invention further discloses an automatic driving method integrating the vehicle-road cooperation and the social network. According to the invention, by integrating the vehicle-road cooperation technology and the social network information, multi-dimensional information interaction between the vehicle and the road infrastructure and between the vehicle and other vehicles is realized, and the accuracy of environment perception and real-time decision making is enhanced.

Description

Technical Field

[0001] The present invention relates to an automatic driving system and method integrating vehicle-road collaboration and social network. Background Art

[0002] With the rapid development of intelligent transportation systems, autonomous driving technology has become an important direction for future transportation. Autonomous driving can not only improve the autonomy and operating efficiency of vehicles, but also significantly reduce the rate of traffic accidents. However, a major challenge facing current autonomous driving technology is how to achieve efficient and safe driving in complex road environments. Relying solely on the vehicle's own sensors and algorithms is difficult to cope with changing road conditions and rapidly changing traffic dynamics. Therefore, solutions based on vehicle-to-everything (V2X) have gradually become the focus.

[0003] Vehicle-road collaboration technology can effectively expand the vehicle's perception range and improve the decision-making ability of the autonomous driving system through information exchange between vehicles and road infrastructure. Specifically, vehicle-road collaboration can realize real-time sharing of traffic signals, road conditions, and other vehicle status, helping autonomous vehicles make more accurate decisions in complex environments. Compared with single-vehicle intelligence, vehicle-road collaboration can further optimize traffic flow, reduce traffic congestion, and improve the safety and reliability of the overall transportation system.

[0004] At the same time, with the popularization of social networks, Internet of Vehicles (IoV) technology has brought new opportunities for autonomous driving. The information exchange between vehicles is not limited to traffic and road conditions. Social networks can also be used to build a collaborative model of "vehicle groups". Through social networks, vehicles can share more dimensional information, such as driving behavior patterns, emergency response, etc., to achieve a higher level of collaborative driving.

[0005] At present, autonomous driving technology has made significant progress, but in complex road environments, the decision-making ability of vehicles is still limited, especially in terms of perception range and handling of emergencies. Single-vehicle intelligent systems mainly rely on on-board sensors, which are difficult to cope with the challenges brought by occlusion or dynamic traffic flow. To this end, vehicle-road collaborative technology expands the vehicle's environmental perception range and improves decision-making efficiency and safety through real-time information interaction between vehicles and road infrastructure. However, data processing and optimization of vehicle-road collaborative technology in practical applications still need to be broken through. At the same time, the development of social networks has provided a new way of information sharing for the Internet of Vehicles. By integrating social networks, vehicles can share more information such as driving behaviors and traffic events, further improving the effect of collaborative driving.

[0006] Therefore, an automatic driving system and method integrating vehicle-road collaboration and social network are provided. Summary of the invention

[0007] The purpose of the present invention is to overcome the existing defects and provide an automatic driving system and method that integrates vehicle-road collaboration and social networks. By integrating vehicle-road collaboration technology and social network information, multi-dimensional information interaction between vehicles and road infrastructure and other vehicles is achieved, thereby enhancing the accuracy of environmental perception and real-time decision-making.

[0008] The technical solution to achieve the above purpose is: An automatic driving system integrating vehicle-road collaboration and social network according to one embodiment of the present invention includes: A data acquisition module is used to collect environmental information data around the vehicle, including the vehicle's own sensor data and external data sources; A communication module, used to transmit the collected environmental information data through cellular vehicle networking technology; The decision-making module is used to conduct comprehensive analysis and decision-making on the collected environmental information data through a deep learning model, accurately predict the current road conditions, and generate a speed control plan; The execution module is used to control the actual movement of the vehicle according to the specified decision output instructions, and the actual movement of the vehicle includes but is not limited to acceleration and deceleration.

[0009] Preferably, in the data acquisition module, the vehicle's own sensors include but are not limited to lidar, cameras, ultrasonic radar and GPS; external data sources include but are not limited to dynamic traffic data and unstructured user-generated data, and social networking platforms are used as acquisition channels.

[0010] Preferably, in the data acquisition module, the data from different sensors are processed through a fusion algorithm and integrated into a consistent environmental model to generate accurate object recognition and positioning information. This is further extended to the communication between the vehicle and the roadside unit through the vehicle-road collaborative perception technology, and the edge computing technology is used to process the perception data at the edge node, so that the vehicle can perceive pedestrians, other vehicles or obstacles in blind spots or obstructions in advance and optimize driving decisions.

[0011] Preferably, the decision module includes: The data fusion unit is used to model the relationship between dynamic traffic data and unstructured user-generated data into a graph structure through a graph neural network, and to pass messages on the graph to achieve dynamic fusion of multi-dimensional data; A feature extraction unit, which is used to extract features of the vehicle's own sensor data through a convolutional neural network and process time series data through a long short-term memory network; The vehicle speed guidance unit is used to establish a maximum green wave bandwidth optimization model based on a universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, obtain a vehicle speed guidance model, and then generate vehicle acceleration and deceleration decisions; The collaborative optimization unit is used to adjust the traffic signal timing according to the signal control, and the vehicle speed guidance model adjusts the vehicle's driving speed.

[0012] Preferably, in the vehicle speed guidance unit, based on the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, the maximum green wave bandwidth optimization model is established by determining the value space of adjacent ideal spacings, deriving the calculation formulas for the uplink and downlink offset green signal ratios, and calculating the formula for the green signal ratios above and below the ideal green light centerline; that is, Ideal spacing and value range of adjacent intersections: ; ; In the formula, is the ideal spacing between adjacent intersections, is the initial speed of the vehicle arriving at the intersection, is the average speed of vehicles traveling at the intersection, Stopping sight distance is the minimum initial velocity of the vehicle arriving at the intersection, is the minimum stopping sight distance, is the maximum value of the initial velocity of the vehicle when it arrives at the intersection, is the average initial velocity of vehicles arriving at the intersection, is the maximum stopping sight distance; The offset green signal ratio refers to the up and down offset of the actual green light midpoint relative to the ideal green light centerline due to the deviation of the actual position of the intersection from the ideal position; for the up and down directions of the main road, the concept of the offset green signal ratio is further expanded to the up and down offset green signal ratios of the main road, that is, the green signal ratios above and below the ideal green light centerline: ; ; In the formula, is the green lane ratio of the main road upward deviation, is the green signal ratio of the trunk road downlink deviation, is the green wave offset; By analyzing the uplink and downlink offset green signal ratios and the green signal ratios above and below the ideal green light centerline and the corresponding calculation methods, it can be obtained that the uplink and downlink green wave bandwidth is the sum of the minimum value of the green signal ratio set above the uplink and downlink ideal green light centerline and the minimum value of the green signal ratio set below it. If the variable Indicates the upstream green wave bandwidth of the trunk road; variable The ideal green light center line represents the green signal ratio above and below the green wave bandwidth of the trunk road, then: ; ; Taking the maximum sum of the two-way green wave bandwidth of the trunk road as the optimization target, the upstream green wave design speed, the downstream green wave design speed and the signal period as the optimization variables, and combining the following formula, the corresponding maximum green wave bandwidth optimization model is established: ; In the formula, is the green wave width, i.e., a time width in the green wave control. Within this time range, vehicles arriving at the green wave control can pass through all intersections in the green wave band smoothly without encountering red lights; Traffic flows consisting of multiple vehicles are called mixed traffic flows. From a longitudinal perspective, mixed traffic flows include artificial following vehicles. , ACC vehicles and CACC vehicles; Assuming that the mixed traffic flow is in equilibrium, the ACC vehicle and the CACC vehicle guide the artificial following vehicle so that the speed of each vehicle in the mixed traffic flow is equal to the equilibrium speed. In the equilibrium state, the distance between the front of the vehicles is equal, and the distance between the front of the vehicles is equal. , and Indicates the headway between manual following vehicles, ACC vehicles and CACC vehicles, with , , Indicates the ratio of manual following vehicles, ACC vehicles and CACC vehicles; Then, let the length of the road section covered by the vehicle be , is the market share of automatic following vehicles, when the total number of vehicles on the road section When it is large enough, the length of the road segment can be regarded as the sum of the headway distances of all vehicles: ; The traffic flow density is the number of vehicles per unit distance in the road section, and the mixed traffic flow density k is: ; when When , the traffic density of mixed traffic flow is: ; when When , the traffic density of mixed traffic flow is: ; According to the above formula, the equilibrium velocity is expressed by density k. : ; In the formula, is the equilibrium velocity, is the minimum safe stopping distance, The safe distance between the front and rear vehicles. is the free flow speed, is the proportion of ACC and CACC vehicles, is the expected inter-vehicle time interval parameter of ACC vehicles, is the expected inter-vehicle time interval parameter of CACC vehicles, For traffic flow.

[0013] Preferably, in the vehicle speed guidance unit, the vehicle speed guidance model includes an acceleration control model, and the objective function of the acceleration control model is that the travel time of the vehicle to the next intersection stop line after the acceleration guidance is the shortest, and its expression is: ; In the formula, is the shortest time required for acceleration, is the actual sight distance at the intersection, The acceleration is the one that a driver generally feels comfortable with; The constraints of the acceleration control model are: ; In the formula, is the green light time within the signal cycle, is the remaining time of the current red light phase, is the current intersection signal cycle, is the next adjacent intersection signal cycle, is the remaining time for the vehicle to travel through the current intersection to the next intersection. The distance from the current intersection's starting stop line to the next intersection's ending stop line. is the initial velocity to the current intersection, is the final speed through the current intersection, Delayed driving time, It is the distance between the stop lines of two adjacent intersections.

[0014] Preferably, in the vehicle speed guidance unit, the vehicle speed guidance model includes a deceleration control model, and the objective function of the deceleration control model optimization is to maximize the speed of the vehicle when it passes through the intersection stop line after deceleration, that is, the time for the vehicle to first decelerate and then travel at a constant speed in the speed control section is: ; In the formula, Actual sight distance at intersection, A deceleration rate that a driver generally feels comfortable with; By rearranging the above formula, the objective function expression of the deceleration control model optimization can be obtained as follows: ; In the formula, The maximum initial speed at which the vehicle can stop at the stop line of the current intersection after deciding that the vehicle cannot pass the next intersection. Stop time for deceleration; The constraints of the deceleration control model are: ; In the formula, is the current intersection signal cycle.

[0015] Preferably, in the collaborative optimization unit, the signal control and vehicle speed guidance are coordinated to optimize the system with a double-layer hierarchical structure, and a double-layer programming model is used to solve the problem; the upper model takes the maximum continuous throughput of the trunk line as the objective function and the guidance speed as the control variable; the lower model takes the weighted average of the minimization of the delay value and the average number of stops as the objective function and the phase difference as the control variable; The upper model uses the guide speed as the decision variable and sets the maximum continuous throughput of the trunk line as the objective function; The rolling time algorithm is used to analyze the traffic volume in the time series and construct the objective function; Through the rolling time algorithm, the traffic volume is counted within a given sliding window, and as the sliding window moves backward, the changes in traffic volume are accurately reflected, providing direction for signal timing adjustment; The optimization objective function of the upper model is to maximize the sum of vehicles that continuously pass through each intersection trunk line, while taking into account the travel time of the vehicles; the objective function of the upper model can be expressed by the following formula: ; In the formula, is the continuous throughput of the trunk line, is the number of intersections involved in decision making, is the initial value of the decision sequence, i.e., the first intersection. is the given number of sliding windows, is the sliding window backward displacement, is the number of intersections, is the intersection number, is the number of decision logic cycles, is the vehicle travel time, is the deceleration braking distance; The constraints of the upper model are the vehicle speed The speed change relationship is satisfied during the guidance process, as shown in the following formula: ; In the formula, is the minimum acceleration / deceleration of the vehicle, is the maximum acceleration / deceleration of the vehicle; The lower-level planning model considers the delay of vehicles in non-coordinated phases and takes the weighted average of the delay value and the average number of stops as the objective function. Comprehensive evaluation index Optimize and provide an optimization solution for the upper model; the expression of the objective function is as follows: ; ; In the formula, is the rolling window ordinal m, is the rolling window ordinal number n, is the sight distance after the intersection decision, is the weight, is the average number of parking times, is an ordinal number; In addition to meeting the minimum green light time and maximum green light time requirements for signal adjustment, the constraints should also control the compressed green light duration of the uncoordinated phase within an acceptable range. The specific constraints are: ; In the formula, Extend the green light time at decision intersections, For acceptable time, Extend the green light time for the actual intersection, For acceptable time, is the actual time when the green light at the intersection turns on. Extended time for decision-making green light, The green light for decision-making comes on early. The minimum time for the green light to be extended or turned on early. The maximum time that the green light can be extended or turned on early.

[0016] A second aspect of the present invention is an autonomous driving method integrating vehicle-road collaboration and social network, comprising: Step S1, collecting environmental information data around the vehicle, including the vehicle's own sensor data and external data; Step S2, comprehensively analyzing and making decisions on the collected environmental information data through a deep learning model, accurately predicting the current road conditions, and generating a speed control plan; Step S3, controlling the actual movement of the vehicle according to the designated decision output instruction, wherein the actual movement of the vehicle is controlled including but not limited to acceleration and deceleration.

[0017] Preferably, in step S1, the vehicle's own sensors include but are not limited to laser radar, camera, ultrasonic radar and GPS; the external data source includes but is not limited to dynamic traffic data and unstructured user-generated data, which is obtained through a social networking platform; The step S2 comprises: Step S21, modeling the relationship between dynamic traffic data and unstructured user-generated data into a graph structure through a graph neural network, and performing message transmission on the graph to achieve dynamic fusion of multi-dimensional data; Step S22, extracting features of the vehicle's own sensor data through a convolutional neural network, and processing the time series data through a long short-term memory network; Step S23, according to the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, a maximum green wave bandwidth optimization model is established to obtain a vehicle speed guidance model, thereby generating a vehicle acceleration and deceleration decision; Step S24, adjusting the traffic signal timing according to the signal control, and adjusting the vehicle speed by the vehicle speed guidance model, thereby generating a speed control plan.

[0018] The beneficial effects of the present invention are as follows: through the combination of vehicle-road collaboration and social networks, the system can obtain more comprehensive environmental information and make up for the shortcomings of single-vehicle perception. The cellular vehicle network and edge computing technology are used to ensure the real-time and reliability of data transmission, meeting the stringent requirements of autonomous driving for efficient communication; at the same time, the system integrates deep learning algorithms, has the ability to predict and respond to complex traffic environments, and improves the accuracy of decision-making; it realizes multi-dimensional information interaction between vehicles and road infrastructure and other vehicles, and enhances the accuracy of environmental perception and real-time decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a module diagram of an automatic driving system integrating vehicle-road collaboration and social network according to the present invention; Figure 2 It is a specific module diagram of the decision module in the present invention; Figure 3 It is a flow chart of an autonomous driving method integrating vehicle-road collaboration and social network according to the present invention; Figure 4 It is a specific flow chart for accurately predicting the current road conditions and generating a speed control scheme in the present invention; Figure 5 It is a judgment flow chart of the lower-level planning model of the present invention when considering the delay of vehicles passing through non-coordinated phases. DETAILED DESCRIPTION

[0020] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance.

[0021] The present invention will be further described below in conjunction with the accompanying drawings.

[0022] like Figure 1 As shown, an autonomous driving system integrating vehicle-road collaboration and social network includes: a data acquisition module 1, a communication module 2, a decision module 3 and an execution module 4.

[0023] The data acquisition module 1 is used to collect environmental information data around the vehicle, including the vehicle's own sensor data and external data sources.

[0024] In the embodiments, the vehicle's own sensors include but are not limited to lidar, cameras, ultrasonic radar and GPS; external data sources include but are not limited to dynamic traffic data and unstructured user-generated data, which are obtained through social networking platforms; for example, drivers' feedback on real-time road conditions, reports of sudden traffic incidents and other driving behavior patterns; social networks can form a "virtual traffic network" through the integration of multi-source data, and transmit key information between vehicles in real time, helping autonomous driving vehicles predict potential dangerous scenarios, traffic jams or emergencies; this information sharing not only helps to improve the efficiency of collaborative driving, but also enhances the overall safety and resilience of the system, and achieves more accurate perception of complex road environments.

[0025] In the embodiment, the data from different sensors are processed through a fusion algorithm and integrated into a consistent environmental model to generate accurate object recognition and positioning information. The data is further extended to the communication between the vehicle and the roadside unit through the vehicle-road collaborative perception technology, and the edge computing technology is used to process the perception data at the edge node, so that the vehicle can perceive pedestrians, other vehicles or obstacles in blind spots or obstructions in advance and optimize driving decisions.

[0026] Communication module 2 is used to transmit the collected environmental information data through cellular vehicle networking technology, and use blockchain technology to provide security for information transmission in the vehicle networking; it supports vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) and vehicle-to-network (V2N) communications; it also introduces edge computing technology to delegate some computing tasks to edge nodes to reduce data transmission delays; at the same time, it uses blockchain technology to ensure the security and credibility of data transmission and prevent data tampering and illegal access; blockchain uses decentralized distributed ledger technology to ensure that the communication between vehicles and infrastructure is credible and prevent information from being tampered with or attacked. At the same time, the system combines homomorphic encryption algorithms to ensure that the privacy information of the vehicle can be protected during data processing and achieve all-round security protection.

[0027] The decision module 3 is used to conduct comprehensive analysis and decision-making on the collected environmental information data through a deep learning model, accurately predict the current road conditions, and generate a speed control plan.

[0028] The execution module 4 is used to control the actual movement of the vehicle according to the specified decision output instructions, and the actual movement of the vehicle includes but is not limited to acceleration and deceleration.

[0029] like Figure 2 As shown, the decision module 3 includes: a data fusion unit 31, a feature extraction unit 32, a vehicle speed guidance unit 33 and a collaborative optimization unit 34.

[0030] The data fusion unit 31 is used to model the relationship between dynamic traffic data and unstructured user-generated data into a graph structure through a graph neural network, and to pass messages on the graph to achieve dynamic fusion of multi-dimensional data; this technology can not only extract useful information from multi-source data, but also help vehicles predict future traffic flow changes by analyzing the historical behavior of other vehicles, thereby further optimizing the decision-making process.

[0031] The feature extraction unit 32 is used to extract the features of the vehicle's own sensor data through a convolutional neural network and process the time series data through a long short-term memory network.

[0032] The present invention uses a deep learning model that combines convolutional neural networks (CNN) and long short-term memory networks (LSTM) to process large amounts of data from the perception layer. CNN is mainly used to extract features of image data, such as real-time road images captured by cameras, while LSTM can process time series data, such as historical data on traffic flow changes. Combining these two technologies, the system can accurately predict current road conditions and generate optimal path planning and speed control solutions.

[0033] The vehicle speed guidance unit 33 is used to establish a maximum green wave bandwidth optimization model according to a universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, obtain a vehicle speed guidance model, and then generate a vehicle acceleration and deceleration decision.

[0034] In the embodiment, based on the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, the value space of adjacent ideal spacings is determined, the calculation formulas for the uplink and downlink offset green signal ratios are derived, and the formula for calculating the green signal ratio above and below the ideal green light centerline is calculated, thereby establishing a maximum green wave bandwidth optimization model. The model takes the maximum sum of the bidirectional green wave bandwidths as the optimization target, and takes the uplink and downlink green wave design speeds and signal periods as optimization variables, overcoming the limitations of the fixed values ​​of the existing numerical solution algorithms and having wide applicability.

[0035] In the embodiment, based on the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, the maximum green wave bandwidth optimization model is established by determining the value space of adjacent ideal spacings, deriving the calculation formulas for the uplink and downlink offset green signal ratios, and calculating the formulas for the green signal ratios above and below the ideal green light centerline; that is, Ideal spacing and value range of adjacent intersections: ; ; In the formula, is the ideal spacing between adjacent intersections, is the initial speed of the vehicle arriving at the intersection, is the average speed of vehicles traveling at the intersection, Stopping sight distance is the minimum initial velocity of the vehicle arriving at the intersection, is the minimum stopping sight distance, is the maximum value of the initial velocity of the vehicle when it arrives at the intersection, is the average initial velocity of vehicles arriving at the intersection, is the maximum stopping sight distance; The offset green signal ratio refers to the up and down offset of the actual green light midpoint relative to the ideal green light centerline due to the deviation of the actual position of the intersection from the ideal position; for the up and down directions of the main road, the concept of the offset green signal ratio is further expanded to the up and down offset green signal ratios of the main road, that is, the green signal ratios above and below the ideal green light centerline: ; ; In the formula, is the green lane ratio of the main road upward deviation, is the green signal ratio of the trunk road downlink deviation, is the green wave offset; By analyzing the uplink and downlink offset green signal ratios and the green signal ratios above and below the ideal green light centerline and the corresponding calculation methods, it can be obtained that the uplink and downlink green wave bandwidth is the sum of the minimum value of the green signal ratio set above the uplink and downlink ideal green light centerline and the minimum value of the green signal ratio set below it. If the variable Indicates the upstream green wave bandwidth of the trunk road; variable The ideal green light center line represents the green signal ratio above and below the green wave bandwidth of the trunk road, then: ; ; Taking the maximum sum of the two-way green wave bandwidth of the trunk road as the optimization target, the upstream green wave design speed, the downstream green wave design speed and the signal period as the optimization variables, and combining the following formula, the corresponding maximum green wave bandwidth optimization model is established: ; In the formula, Green wave width, that is, a time width in green wave control. Within this time range, vehicles arriving at the green wave control can pass through all intersections in the green wave belt smoothly without encountering red lights. Traffic flows consisting of multiple vehicles are called mixed traffic flows. From a longitudinal perspective, mixed traffic flows include artificial following vehicles. , ACC vehicles and CACC vehicles; Assuming that the mixed traffic flow is in equilibrium, the ACC vehicle and the CACC vehicle guide the artificial following vehicle so that the speed of each vehicle in the mixed traffic flow is equal to the equilibrium speed. In the equilibrium state, the distance between the front of the vehicles is equal, and the distance between the front of the vehicles is equal. , and Indicates the headway between manual following vehicles, ACC vehicles and CACC vehicles, with , , Indicates the ratio of manual following vehicles, ACC vehicles and CACC vehicles; Then, let the length of the road section covered by the vehicle be , is the market share of automatic following vehicles, when the total number of vehicles on the road section When it is large enough, the length of the road segment can be regarded as the sum of the headway distances of all vehicles: ; The traffic flow density is the number of vehicles per unit distance in the road section, and the mixed traffic flow density k is: ; when When , the traffic density of mixed traffic flow is: ; when When , the traffic density of mixed traffic flow is: ; According to the above formula, the equilibrium velocity is expressed by density k. : ; In the formula, is the equilibrium velocity, is the minimum safe stopping distance, The safe distance between the front and rear vehicles. is the free flow speed, is the proportion of ACC and CACC vehicles, is the expected inter-vehicle time interval parameter of ACC vehicles, is the expected inter-vehicle time interval parameter of CACC vehicles, For traffic flow.

[0036] In the embodiment, the vehicle speed guidance model includes an acceleration control model. A vehicle starting at time 0 with an initial speed immediately obtains the state of the next intersection signal light after entering the road section. If the straight phase of the current intersection signal is a red light, the remaining time of the current red light phase is obtained. , determine whether the vehicle entering the road section at the initial speed can reach the stop line of the intersection before the end of the next phase green light time at this speed. If not, but the vehicle can reach the stop line of the next intersection within the next phase green light time after acceleration, then the red light acceleration induction strategy is enabled: select an acceleration that drivers generally feel comfortable with , guide the vehicle to accelerate to the maximum speed limit of the road section, and drive at this speed until passing the next intersection; then The objective function of the acceleration control model is to minimize the travel time of the vehicle to the next intersection stop line after acceleration guidance, and its expression is: ; ; In the formula, is the shortest time required for acceleration, The actual sight distance of the intersection The acceleration is the one that a driver generally feels comfortable with; The constraints of the acceleration control model are: ; In the formula, is the green light time within the signal cycle, is the remaining time of the current red light phase, is the current intersection signal cycle, is the next adjacent intersection signal cycle, is the remaining time for the vehicle to travel through the current intersection to the next intersection. The distance from the current intersection's starting stop line to the next intersection's ending stop line. is the initial velocity to the current intersection, is the final speed through the current intersection, Delayed driving time, It is the distance between the stop lines of two adjacent intersections.

[0037] In the embodiment, the vehicle speed guidance model includes a deceleration control model. A vehicle starting at time 0 with an initial speed immediately obtains the state of the next intersection signal light after entering the road section. If the straight-ahead phase of the current intersection signal is a red light, the remaining time of the current red light phase is obtained to determine whether the vehicle entering the road section at the initial speed can reach the intersection stop line within the green light time of the next phase at this speed. If not, it is determined whether the vehicle entering the road section at the initial speed can reach the intersection stop line within the green light time of the next phase by decelerating. If so, the red light deceleration induction strategy is enabled: a deceleration that drivers generally feel comfortable with is selected to guide the vehicle to decelerate, so that the vehicle starts to decelerate at an appropriate position and can reach the intersection stop line within the green light time of the next cycle, thereby avoiding the vehicle stopping and waiting. The objective function of the deceleration control model optimization is to maximize the speed of the vehicle when it passes the intersection stop line after deceleration, that is, the time it takes for the vehicle to decelerate and then travel at a constant speed in the speed control section is: ; In the formula, Actual sight distance at intersection A deceleration rate that a driver generally feels comfortable with; By rearranging the above formula, the objective function expression of the deceleration control model optimization can be obtained as follows: ; ; In the formula, The maximum initial speed at which the vehicle can stop at the stop line of the current intersection after deciding that the vehicle cannot pass the next intersection. Stop time for deceleration; The constraints of the deceleration control model are: ; In the formula, is the current intersection signal cycle.

[0038] The collaborative optimization unit 34 is used to adjust the traffic signal timing according to the signal control, and the vehicle speed guidance model adjusts the vehicle's driving speed.

[0039] In the embodiment, in the urban traffic system, signal control and vehicle speed guidance are two key trunk control strategies. Signal control adjusts the timing of traffic signals, while vehicle speed guidance reduces traffic delays by adjusting the speed of vehicles; signal control and vehicle speed guidance cooperate with the system optimization with a two-layer hierarchical structure, and are solved using a two-layer programming model; the upper model takes the maximum continuous throughput of the trunk as the objective function and the guidance speed as the control variable; the lower model takes the weighted average of the minimization of the delay value and the average number of stops as the objective function and the phase difference as the control variable; The upper model uses the guide speed as the decision variable and sets the maximum continuous throughput of the trunk line as the objective function; The rolling time algorithm is used to analyze the traffic volume in the time series and construct the objective function; Through the rolling time algorithm, the traffic volume is counted within a given sliding window, and as the sliding window moves backward, the changes in traffic volume are accurately reflected, providing direction for signal timing adjustment; The optimization objective function of the upper model is to maximize the sum of vehicles that continuously pass through each intersection trunk line, while taking into account the travel time of the vehicles; the objective function of the upper model can be expressed by the following formula: ; In the formula, is the continuous throughput of the trunk line, is the number of intersections involved in decision making, is the initial value of the decision sequence, i.e., the first intersection. is the given number of sliding windows, is the sliding window backward displacement, is the number of intersections, is the intersection number, is the number of decision logic cycles, is the vehicle travel time, is the deceleration braking distance; The constraints of the upper model are the vehicle speed The speed change relationship is satisfied during the guidance process, as shown in the following formula: ; In the formula, is the minimum acceleration / deceleration of the vehicle, is the maximum acceleration / deceleration of the vehicle; The lower-level planning model considers the delay of vehicles in non-coordinated phases and takes the weighted average of the delay value and the average number of stops as the objective function. Comprehensive evaluation index Optimize and provide an optimization solution for the upper model; the expression of the objective function is as follows: ; ; In the formula, is the rolling window ordinal m, is the rolling window ordinal number n, is the sight distance after the intersection decision, is the weight, is the average number of parking times, is an ordinal number; In addition to meeting the minimum green light time and maximum green light time requirements for signal adjustment, the constraints should also control the compressed green light duration of the uncoordinated phase within an acceptable range. The specific constraints are: ; In the formula, Extend the green light time at decision intersections, For acceptable time, Extend the green light time for the actual intersection, For acceptable time, is the actual time when the green light at the intersection turns on. Extended time for decision-making green light, The green light for decision-making comes on early. The minimum time for the green light to be extended or turned on early. The maximum time that the green light can be extended or turned on early.

[0040] like Figure 3 As shown, an autonomous driving method integrating vehicle-road collaboration and social network includes: Step S1, collecting environmental information data around the vehicle, including the vehicle's own sensor data and external data.

[0041] In the embodiment, the vehicle's own sensors include but are not limited to lidar, cameras, ultrasonic radar and GPS; external data sources include but are not limited to dynamic traffic data and unstructured user-generated data, which are obtained through social networking platforms.

[0042] Step S2, using a deep learning model to comprehensively analyze and make decisions on the collected environmental information data, accurately predict the current road conditions, and generate a speed control plan.

[0043] like Figure 4 As shown, step S2 includes: Step S21, the relationship between dynamic traffic data and unstructured user-generated data is modeled as a graph structure through a graph neural network, and message passing is performed on the graph to achieve dynamic fusion of multi-dimensional data.

[0044] Step S22, extracting the features of the vehicle's own sensor data through a convolutional neural network, and processing the time series data through a long short-term memory network.

[0045] Step S23, based on the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, a maximum green wave bandwidth optimization model is established to obtain a vehicle speed guidance model, thereby generating a vehicle acceleration and deceleration decision.

[0046] Step S24, adjusting the traffic signal timing according to the signal control, and adjusting the vehicle speed by the vehicle speed guidance model, thereby generating a speed control plan.

[0047] Step S3, controlling the actual movement of the vehicle according to the designated decision output instruction, wherein the actual movement of the vehicle is controlled including but not limited to acceleration and deceleration.

[0048] The lower-level planning model considers the delay of vehicles in non-coordinated phases, for example Figure 5 As shown, when the straight-ahead phase is green when the vehicle enters the speed guidance area, it is determined whether the straight-ahead phase is green when the vehicle arrives at the intersection. If it is green, the original plan remains unchanged; when the straight-ahead phase is red when the vehicle enters the speed guidance area, it is determined whether the straight-ahead phase is red when the vehicle arrives at the intersection. If it is not red, the original plan remains unchanged.

[0049] The phase compression method proposed in this invention has a certain impact on the delay of vehicles in non-coordinated phases, but the increased delay level is relatively small and does not have a large fluctuation impact on vehicles in non-coordinated phases. At the same time, the reduction in the overall delay level of the trunk line shows that the overall benefits of this method are better than the adverse effects, and the overall traffic efficiency of the trunk line traffic flow is improved.

[0050] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments may still be modified, or some or all of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An autonomous driving system integrating vehicle-road collaboration and social network, characterized in that: include: A data acquisition module is used to collect environmental information data around the vehicle, including the vehicle's own sensor data and external data sources; A communication module, used to transmit the collected environmental information data through cellular vehicle networking technology; The decision-making module is used to conduct comprehensive analysis and decision-making on the collected environmental information data through a deep learning model, accurately predict the current road conditions, and generate a speed control plan; The execution module is used to control the actual movement of the vehicle according to the specified decision output instructions, and the actual movement of the vehicle includes but is not limited to acceleration and deceleration.

2. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 1, characterized in that: In the data acquisition module, the vehicle's own sensors include but are not limited to lidar, cameras, ultrasonic radar and GPS; external data sources include but are not limited to dynamic traffic data and unstructured user-generated data, which are obtained through social networking platforms.

3. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 1, characterized in that: In the data acquisition module, the data from different sensors are processed through a fusion algorithm and integrated into a consistent environmental model to generate accurate object recognition and positioning information. The vehicle-road collaborative perception technology is further extended to the communication between the vehicle and the roadside unit, and the edge computing technology is used to process the perception data at the edge node, so that the vehicle can perceive pedestrians, other vehicles or obstacles in blind spots or obstructions in advance and optimize driving decisions.

4. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 1, characterized in that: The decision module comprises: The data fusion unit is used to model the relationship between dynamic traffic data and unstructured user-generated data into a graph structure through a graph neural network, and to pass messages on the graph to achieve dynamic fusion of multi-dimensional data; A feature extraction unit, which is used to extract features of the vehicle's own sensor data through a convolutional neural network and process time series data through a long short-term memory network; The vehicle speed guidance unit is used to establish a maximum green wave bandwidth optimization model based on a universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, obtain a vehicle speed guidance model, and then generate vehicle acceleration and deceleration decisions; The collaborative optimization unit is used to adjust the traffic signal timing according to the signal control, and the vehicle speed guidance model adjusts the vehicle's driving speed.

5. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 4, characterized in that: In the vehicle speed guidance unit, based on the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, the maximum green wave bandwidth optimization model is established by determining the value space of adjacent ideal spacings, deriving the calculation formulas for the uplink and downlink offset green signal ratios, and calculating the formula for the green signal ratios above and below the ideal green light centerline; that is, Ideal spacing and value range of adjacent intersections: ； ； In the formula, is the ideal spacing between adjacent intersections, is the initial speed of the vehicle arriving at the intersection, is the average speed of vehicles traveling at the intersection, is the stopping sight distance, is the minimum initial speed of the vehicle arriving at the intersection, is the minimum stopping sight distance, is the maximum value of the initial velocity of the vehicle when it arrives at the intersection, is the average initial velocity of vehicles arriving at the intersection, is the maximum stopping sight distance; The offset green signal ratio refers to the up and down offset of the actual green light midpoint relative to the ideal green light centerline due to the deviation of the actual position of the intersection from the ideal position; for the up and down directions of the main road, the concept of the offset green signal ratio is further expanded to the up and down offset green signal ratios of the main road, that is, the green signal ratios above and below the ideal green light centerline: ； ； In the formula, is the green lane ratio of the main road upward deviation, is the green signal ratio of the trunk road downlink deviation, is the green wave offset; By analyzing the uplink and downlink offset green signal ratios and the green signal ratios above and below the ideal green light centerline and the corresponding calculation methods, it can be obtained that the uplink and downlink green wave bandwidth is the sum of the minimum value of the green signal ratio set above the uplink and downlink ideal green light centerline and the minimum value of the green signal ratio set below it. If the variable Indicates the upstream green wave bandwidth of the trunk road; variable The ideal green light center line represents the green signal ratio above and below the green wave bandwidth of the trunk road, then: ； ； Taking the maximum sum of the two-way green wave bandwidth of the trunk road as the optimization target, the upstream green wave design speed, the downstream green wave design speed and the signal period as the optimization variables, and combining the following formula, the corresponding maximum green wave bandwidth optimization model is established: ； In the formula, is the green wave band width, i.e., a time width in the green wave control. Within this time range, vehicles arriving at the green wave control can pass through all intersections in the green wave band smoothly without encountering red lights; Traffic flows consisting of multiple vehicles are called mixed traffic flows. From a longitudinal perspective, mixed traffic flows include artificial following vehicles. , ACC vehicles and CACC vehicles; Assuming that the mixed traffic flow is in equilibrium, the ACC vehicle and the CACC vehicle guide the artificial following vehicle so that the speed of each vehicle in the mixed traffic flow is equal to the equilibrium speed. In the equilibrium state, the distance between the front of the vehicles is equal, and the distance between the front of the vehicles is equal. , and Indicates the headway between manual following vehicles, ACC vehicles and CACC vehicles, with , , Indicates the ratio of manual following vehicles, ACC vehicles and CACC vehicles; Then, let the length of the road section covered by the vehicle be , is the market share of automatic following vehicles, when the total number of vehicles on the road section When it is large enough, the length of the road segment can be regarded as the sum of the headway distances of all vehicles: ； The traffic flow density is the number of vehicles per unit distance in the road section, and the mixed traffic flow density k is: ； when When , the traffic density of mixed traffic flow is: ； when When , the traffic density of mixed traffic flow is: ； According to the above formula, the equilibrium velocity is expressed by density k. : ； In the formula, is the equilibrium velocity, is the minimum safe stopping distance, The safe distance between the front and rear vehicles. is the free flow speed, is the proportion of ACC and CACC vehicles, is the expected inter-vehicle time interval parameter of ACC vehicles, is the expected inter-vehicle time interval parameter of CACC vehicles, For traffic flow.

6. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 5, characterized in that: In the vehicle speed guidance unit, the vehicle speed guidance model includes an acceleration control model. The objective function of the acceleration control model is to minimize the travel time of the vehicle to the next intersection stop line after the acceleration guidance. The expression is: ； ； In the formula, is the shortest time required for acceleration, is the actual sight distance at the intersection, The acceleration is the one that a driver generally feels comfortable with; The constraints of the acceleration control model are: ； In the formula, is the green light time within the signal cycle, is the remaining time of the current red light phase, is the current intersection signal cycle, is the next adjacent intersection signal cycle, is the remaining time for the vehicle to travel through the current intersection to the next intersection. The distance from the current intersection's starting stop line to the next intersection's ending stop line. is the initial velocity to the current intersection, is the final speed through the current intersection, Delayed driving time, It is the distance between the stop lines of two adjacent intersections.

7. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 6, characterized in that: In the vehicle speed guidance unit, the vehicle speed guidance model includes a deceleration control model. The objective function of the deceleration control model optimization is to maximize the speed of the vehicle when it passes through the intersection stop line after deceleration, that is, the time for the vehicle to decelerate first and then travel at a constant speed in the speed control section is: ； In the formula, Actual sight distance at intersection, A deceleration rate that a driver generally feels comfortable with; By rearranging the above formula, the objective function expression of the deceleration control model optimization can be obtained as follows: ； ； In the formula, The maximum initial speed at which the vehicle can stop at the stop line of the current intersection after deciding that the vehicle cannot pass the next intersection. Stop time for deceleration; The constraints of the deceleration control model are: ； In the formula, is the current intersection signal cycle.

8. The autonomous driving system integrating vehicle-road collaboration and social network according to claim 7, characterized in that: In the collaborative optimization unit, the signal control and vehicle speed guidance are coordinated to optimize the system with a double-layer hierarchical structure, and are solved by a double-layer programming model; the upper model takes the maximum continuous throughput of the trunk line as the objective function and the guidance speed as the control variable; the lower model takes the minimization of the weighted average of the delay value and the average number of stops as the objective function and the phase difference as the control variable; The upper model uses the guide speed as the decision variable and sets the maximum continuous throughput of the trunk line as the objective function; The rolling time algorithm is used to analyze the traffic volume in the time series and construct the objective function; Through the rolling time algorithm, the traffic volume is counted within a given sliding window, and as the sliding window moves backward, the changes in traffic volume are accurately reflected, providing direction for signal timing adjustment; The optimization objective function of the upper model is to maximize the sum of vehicles that continuously pass through each intersection trunk line, while taking into account the travel time of the vehicles; the objective function of the upper model can be expressed by the following formula: ； In the formula, is the continuous throughput of the trunk line, is the number of intersections involved in decision making, is the initial value of the decision sequence, i.e., the first intersection. is the given number of sliding windows, is the sliding window backward displacement, is the number of intersections, is the intersection number, is the number of decision logic cycles, is the vehicle travel time, is the deceleration braking distance; The constraints of the upper model are the vehicle speed The speed change relationship is satisfied during the guidance process, as shown in the following formula: ； In the formula, is the minimum acceleration / deceleration of the vehicle, is the maximum acceleration / deceleration of the vehicle; The lower-level planning model considers the delay of vehicles in non-coordinated phases and takes the weighted average of the delay value and the average number of stops as the objective function. Comprehensive evaluation index Optimize and provide an optimization solution for the upper model; the expression of the objective function is as follows: ； ； In the formula, is the rolling window ordinal m, is the rolling window ordinal number n, is the sight distance after the intersection decision, is the weight, is the average number of parking times, is an ordinal number; In addition to meeting the minimum green light time and maximum green light time requirements for signal adjustment, the constraints should also control the compressed green light duration of the uncoordinated phase within an acceptable range. The specific constraints are: ； In the formula, Extend the green light time at decision intersections, For acceptable time, Extend the green light time for the actual intersection, For acceptable time, is the actual time when the green light at the intersection turns on. Extended time for decision-making green light, The green light for decision-making comes on early. The minimum time for the green light to be extended or turned on early. The maximum time that the green light can be extended or turned on early.

9. An automatic driving method based on the automatic driving system integrating vehicle-road collaboration and social network according to claim 1, characterized in that: include: Step S1, collecting environmental information data around the vehicle, including the vehicle's own sensor data and external data; Step S2, comprehensively analyzing and making decisions on the collected environmental information data through a deep learning model, accurately predicting the current road conditions, and generating a speed control plan; Step S3, controlling the actual movement of the vehicle according to the designated decision output instruction, wherein the actual movement of the vehicle is controlled including but not limited to acceleration and deceleration.

10. The method for autonomous driving integrating vehicle-road collaboration and social network according to claim 9, characterized in that: In step S1, the vehicle's own sensors include but are not limited to laser radar, camera, ultrasonic radar and GPS; external data sources include but are not limited to dynamic traffic data and unstructured user-generated data, which are obtained through social network platforms; The step S2 comprises: Step S21, modeling the relationship between dynamic traffic data and unstructured user-generated data into a graph structure through a graph neural network, and performing message transmission on the graph to achieve dynamic fusion of multi-dimensional data; Step S22, extracting features of the vehicle's own sensor data through a convolutional neural network, and processing the time series data through a long short-term memory network; Step S23, according to the universal numerical solution algorithm for the maximum bidirectional green wave bandwidth, a maximum green wave bandwidth optimization model is established to obtain a vehicle speed guidance model, thereby generating a vehicle acceleration and deceleration decision; Step S24, adjusting the traffic signal timing according to the signal control, and adjusting the vehicle speed by the vehicle speed guidance model, thereby generating a speed control plan.

Images