A road information collection data analysis method and system based on artificial intelligence

Abstract

The present invention discloses a road information collection data analysis method and system based on artificial intelligence, which relates to the field of road monitoring technology, including deploying sensors to collect road data, determining the road driving direction, installing independent controllers for each road driving direction to optimize traffic signals; analyzing traffic accidents based on the collected road data, and judging the road passability of traffic accidents, and controlling traffic signals based on the road passability; the control center remotely monitors road data and traffic accidents, and responds and stores them in real time. The present invention effectively improves the traffic efficiency of intersections by independently controlling the road driving direction and comprehensively optimizing the traffic signals of the driving direction, and adjusts the road traffic strategy by detecting the passability of road traffic accidents, thereby improving the safety and flexibility of road driving and reducing the possibility of secondary injuries caused by traffic accidents.

Description

Technical Field

[0001] The present invention relates to the technical field of road monitoring, and in particular to an artificial intelligence-based road information collection data analysis method and system. Background Art

[0002] With the acceleration of modern urbanization, traffic congestion is becoming increasingly serious worldwide. Traditional traffic management methods, such as timed signal light control and traffic management methods that rely on manual adjustment, can no longer meet the increasingly complex traffic needs. To meet this challenge, the Intelligent Transportation System (ITS) came into being. ITS integrates sensors, communication technology and computing technology to monitor and manage traffic flow in real time to improve road efficiency and driving safety. However, existing intelligent transportation systems usually rely on fixed sensor networks to collect traffic data and adjust traffic signals as a whole. They cannot be optimized individually in a certain direction, and lack flexible and effective means for detecting and responding to traffic accidents. Summary of the invention

[0003] In view of the problems existing in the above-mentioned existing road information collection data analysis method and system based on artificial intelligence, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that the existing intelligent transportation system usually relies on a fixed sensor network to collect traffic data and adjust traffic signals as a whole, and cannot be optimized separately for a certain direction, and lacks flexible and effective means for detecting and responding to traffic accidents.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for analyzing road information collection data based on artificial intelligence, which includes deploying sensors to collect road data, determining the road driving direction, and installing an independent controller for each road driving direction to optimize traffic signals;

[0006] Analyze traffic accidents based on collected road data, determine the road passability of the traffic accident, and control traffic signals based on the road passability;

[0007] The control center remotely monitors road data and traffic accidents, and responds and stores them in real time.

[0008] As a preferred solution of the road information collection and data analysis method based on artificial intelligence of the present invention, wherein: the deployment of sensors to collect road data and determine the road driving direction, and the installation of independent controllers for each road driving direction to optimize traffic signals refers to the deployment of sensors on the road to collect road data and pre-process the road data, and determine the driving direction according to the road data;

[0009] Install independent controllers for traffic lights in each driving direction, and connect independent controllers at the same intersection through the network;

[0010] Calculate the cumulative arrival volume A(t) and cumulative departure volume D(t) of peak-hour traffic in each driving direction:

[0011]

[0012]

[0013] Where t is the cumulative time, For time The vehicle arrival rate, For time The vehicle departure rate;

[0014] Calculate the average rush hour delay d(t) for each direction of travel:

[0015]

[0016] Where c is the traffic signal cycle, g is the green light time, x(t) is the saturated traffic flow at time t, q(t) is the vehicle arrival rate at time t, and S is the saturated vehicle arrival rate;

[0017] Analyze the peak-hour queue length Q(t) for each driving direction based on road data:

[0018] The independent controllers of each driving direction at the same intersection are used as nodes, and the influence between driving directions is used as edges to construct a graph neural network for the same intersection.

[0019] , if there is an influence between the driving directions between nodes, edge connections are constructed, and the cumulative traffic arrival volume A(t), cumulative departure volume D(t), average delay d(t) and queue length Q(t) during peak hours of each driving direction are calculated, and recorded in the independent controller of each driving direction as node features;

[0020] Extract the traffic flow in each driving direction and calculate the mean of the traffic flow between two connected nodes in the graph neural network as the weight of the edge connection;

[0021] The graph neural network updates the hidden state of each node by aggregating adjacent nodes and edge connections, uses the back-propagation algorithm to train the graph neural network, defines the minimization loss function to optimize the node state, and uses the reinforcement learning Q-learnning algorithm to optimize the node features based on the optimized node state;

[0022] The traffic signal cycle c, green light time g, vehicle arrival rate q(t) at time t, and saturated vehicle arrival rate S are taken as combinations, and all possible combinations are analyzed as the initial population according to the optimized node characteristics;

[0023] Calculate the fitness of each combination in the initial population, select the combination with the highest fitness as the next generation parent combination, generate new combinations from the parent combinations through crossover and mutation and add them to the initial population;

[0024] Iteratively repeat the selection, crossover and mutation operations until the preset number of iterations is reached or the fitness begins to converge, and then stop the iteration. The combination with the highest fitness in the population formed by stopping the iteration is taken as the optimal combination, and the traffic signal cycle c and green light time g in the optimal combination are extracted as the traffic signal control strategy for each driving direction.

[0025] As a preferred solution of the road information collection data analysis method based on artificial intelligence of the present invention, wherein: the analysis of traffic accidents based on the collected road data includes:

[0026] The YOLOv4 algorithm is used to detect vehicles, pedestrians and obstacles on the road in real time, and output a detection frame containing the object category and location;

[0027] Define vehicles and pedestrians as traffic participants, and use convolutional neural networks to analyze traffic participant images to detect traffic accidents;

[0028] If a traffic accident is detected, the location of the traffic accident is determined through GPS positioning, and emergency personnel are notified to handle the traffic accident. The location of the intersection around the traffic accident is simultaneously determined through GPS positioning, and the independent controllers of each driving direction in the intersection are connected as points to form a rectangle. The minimum rectangle containing the location of the traffic accident is determined, and the independent controllers and driving directions in the minimum rectangle are marked.

[0029] As a preferred solution of the road information collection and data analysis method based on artificial intelligence of the present invention, wherein: the judging of the road passability of the traffic accident and the controlling of the traffic signal based on the road passability refer to using the DeepLabv3+ algorithm to perform semantic segmentation on the traffic accident road after detecting the traffic accident, and identifying road elements including lane lines, sidewalk lines and green belts;

[0030] Traffic accidents are detected using the YOLOv4 algorithm, marked using detection boxes, and identified using lane lines as lanes occupied and lanes not occupied by traffic accidents. If there is an unoccupied lane, the traffic accident is judged to be passable; if there is no unoccupied lane, the traffic accident is judged to be impassable.

[0031] If the traffic accident makes the road impassable, the traffic signal in the marked direction of the intersection and the independent controller will be controlled to be a red light to stop vehicles from passing, and the red light in the driving direction will be controlled to flash continuously to remind vehicles that there is a traffic accident and they need to detour;

[0032] If the traffic accident is passable, the number of unoccupied lanes on the road is identified, and the proportion of unoccupied lanes to the total lanes on the road is calculated. The cumulative traffic arrival volume A(t) and the cumulative departure volume D(t) during peak hours are reduced proportionally. At the same time, the average delay d(t) and the queue length Q(t) are increased proportionally. The traffic signal cycle c and the green light time g are then substituted into the graph neural network for optimization and determined as the new traffic signal control strategy and implemented.

[0033] As a preferred solution of the road information collection and data analysis method based on artificial intelligence described in the present invention, the sensors include laser radar, camera, radar sensor, traffic flow sensor and GPS sensor.

[0034] As a preferred solution of the road information collection and data analysis method based on artificial intelligence described in the present invention, wherein: the control center remotely monitors road data and traffic accidents means that the control center remotely receives road data and traffic accident analysis results collected by sensors and independent controllers through the network, and forms road data into visual charts for real-time display, and simultaneously manually analyzes the severity of the traffic accident according to the location of the traffic accident and the traffic accident image, and notifies relevant personnel to handle the traffic accident.

[0035] As a preferred solution of the road information collection and data analysis method based on artificial intelligence described in the present invention, the real-time response and storage refers to the real-time storage of road data and traffic accident analysis results received by the control center through a database, synchronously responding to the traffic signal control strategy formulated by the independent controller of each driving direction at the intersection, and binding the traffic signal control strategy with the positioning of the independent controller and storing them in a classified manner.

[0036] Another object of the present invention is to provide a road information collection data analysis system based on artificial intelligence, which includes a data collection module for deploying sensors to collect road data and perform preprocessing;

[0037] A traffic control module, which is used to optimize the traffic signal for each driving direction at the intersection based on the collected road data;

[0038] An accident detection module is used to detect traffic accidents on the road, analyze the passability of the traffic accidents, and control traffic signals based on the passability of the traffic accidents;

[0039] The monitoring and storage module is used to remotely receive road data and traffic accident analysis results, and store them in a classified manner through a database.

[0040] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned road information collection data analysis method based on artificial intelligence when executing the computer program.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned road information collection data analysis method based on artificial intelligence.

[0042] The beneficial effects of the present invention are as follows: the present invention effectively improves the traffic efficiency of intersections by independently controlling the road driving direction and comprehensively optimizing the traffic signals in the driving direction, and adjusts the road traffic strategy by detecting the passability of road traffic accidents, thereby improving the safety and flexibility of road driving and reducing the possibility of secondary injuries caused by traffic accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0044] Figure 1 The figure is a flowchart of the road information collection data analysis method based on artificial intelligence.

[0045] Figure 2 Schematic diagram of the flow of traffic signal optimization for a stand-alone controller.

[0046] Figure 3 This is a structural diagram of the road information collection and data analysis system based on artificial intelligence. DETAILED DESCRIPTION

[0047] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0048] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0049] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0050] Example 1

[0051] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and which provides a road information collection data analysis method based on artificial intelligence. The road information collection data analysis method based on artificial intelligence includes:

[0052] S1. Deploy sensors to collect road data, determine the road driving direction, and install independent controllers for each road driving direction to optimize traffic signals;

[0053] Specifically, sensors are deployed to collect road data, determine the road driving direction, and independent controllers are installed in each road driving direction. Traffic signal optimization refers to collecting road data by deploying sensors on the road and pre-processing the road data, and determining the driving direction according to the road data;

[0054] Install independent controllers for traffic lights in each driving direction, and connect independent controllers at the same intersection through the network;

[0055] Calculate the cumulative arrival volume A(t) and cumulative departure volume D(t) of peak-hour traffic in each driving direction:

[0056]

[0057]

[0058] Where t is the cumulative time, For time The vehicle arrival rate, For time The vehicle departure rate is obtained through traffic flow sensors;

[0059] Calculate the average rush hour delay d(t) for each direction of travel:

[0060]

[0061] Where c is the traffic signal cycle, g is the green light time, x(t) is the saturated vehicle flow at time t, q(t) is the vehicle arrival rate at time t, and S is the saturated vehicle arrival rate, which is obtained by the traffic flow sensor;

[0062] Analyze the peak-hour queue length Q(t) for each driving direction based on road data:

[0063] The independent controllers of each driving direction at the same intersection are used as nodes, and the influence between driving directions is used as edges to construct a graph neural network for the same intersection.

[0064] , if there is an influence between the driving directions of the nodes, an edge connection is constructed. For example, if the straight driving direction and the left and right turning driving directions on the same side of the intersection do not influence each other, then there is no edge connection between the nodes of the straight driving direction and the left and right turning driving directions. If there is a mutual influence between the left turning driving direction on one side of the intersection and the straight driving direction on the opposite side, then there is an edge connection between the left turning driving direction on one side of the intersection and the straight driving direction on the opposite side. The cumulative traffic arrival volume A(t), cumulative departure volume D(t), average delay d(t) and queue length Q(t) during peak hours of each driving direction are calculated and recorded in the independent controller of each driving direction as node features;

[0065] Extract the traffic flow in each driving direction and calculate the mean of the traffic flow between two connected nodes in the graph neural network as the weight of the edge connection;

[0066] The graph neural network updates the hidden state of each node by aggregating adjacent nodes and edge connections, uses the back-propagation algorithm to train the graph neural network, defines the minimization loss function to optimize the node state, and uses the reinforcement learning Q-learnning algorithm to optimize the node features based on the optimized node state;

[0067] The traffic signal cycle c, green light time g, vehicle arrival rate q(t) at time t, and saturated vehicle arrival rate S are taken as combinations, and all possible combinations are analyzed as the initial population according to the optimized node characteristics;

[0068] Calculate the fitness of each combination in the initial population, select the combination with the highest fitness as the next generation parent combination, generate new combinations from the parent combinations through crossover and mutation and add them to the initial population;

[0069] Iteratively repeat the selection, crossover and mutation operations until the preset number of iterations is reached or the fitness begins to converge, and then stop the iteration. The combination with the highest fitness in the population formed by stopping the iteration is taken as the optimal combination, and the traffic signal cycle c and green light time g in the optimal combination are extracted as the traffic signal control strategy for each driving direction.

[0070] In traditional traffic signal control systems, the cycle and phase of traffic lights are usually preset, and the cycle and phase of traffic lights are not adjusted according to actual traffic conditions, resulting in unreasonable signal light switching times during peak and non-peak hours, and unable to adapt to dynamically changing traffic flows. Instead, accurate mathematical models are used to calculate the cumulative arrivals, cumulative departures, average delays, and queue lengths during peak hours, and to dynamically adjust the cycle and green light time of traffic lights to ensure that the signal control strategy matches the actual traffic conditions. By deploying sensors on the road, road data can be collected in real time and accurately, including vehicle arrival and departure rates, traffic flows, etc. After preprocessing, these data can be It can effectively filter noise and outliers to ensure the accuracy and reliability of data, install independent controllers in each driving direction, and connect these controllers through the network to achieve independent control of traffic lights in each direction, improve the flexibility and response speed of signal control, and dynamically adjust the duration and phase of traffic lights according to real-time traffic conditions to reduce traffic congestion. It uses graph neural networks and reinforcement learning algorithms to update node states by aggregating adjacent nodes and edge connections, minimize loss functions, optimize node features, and then optimize traffic signal control strategies, further improving the intelligence level and optimization effect of traffic signal control, and ensuring dynamic adjustment and optimal control of signal control strategies.

[0071] Furthermore, the sensors include lidar, camera, radar sensor, traffic flow sensor and GPS sensor.

[0072] S2. Analyze traffic accidents based on the collected road data, determine the road passability of the traffic accident, and control traffic signals based on the road passability;

[0073] Specifically, traffic accidents analyzed based on collected road data include:

[0074] The YOLOv4 algorithm is used to detect vehicles, pedestrians and obstacles on the road in real time, and output a detection frame containing the object category and location;

[0075] Define vehicles and pedestrians as traffic participants, and use convolutional neural networks to analyze traffic participant images to detect traffic accidents;

[0076] If a traffic accident is detected, the location of the traffic accident is determined through GPS positioning, and emergency personnel are notified to handle the traffic accident. The location of the intersection around the traffic accident is simultaneously determined through GPS positioning, and the independent controllers of each driving direction in the intersection are connected as points to form a rectangle. The minimum rectangle containing the location of the traffic accident is determined, and the independent controllers and driving directions in the minimum rectangle are marked.

[0077] The efficient detection capability of the YOLOv4 algorithm ensures the real-time discovery of traffic accidents and significantly improves the response speed of traffic accidents. Compared with traditional manual monitoring systems, YOLOv4 can complete detection within milliseconds, greatly shortening the accident discovery and response time. The automated target detection process does not require manual intervention and can work continuously around the clock, improving the coverage and work efficiency of traffic monitoring. CNN can automatically extract image features and classify them, accurately identify traffic accident scenes, and improve the intelligence level of accident detection. Through real-time image analysis, CNN can make quick judgments after the accident and trigger the emergency response mechanism in time. GPS positioning technology can quickly and accurately locate the location of the traffic accident, ensuring that emergency personnel and relevant departments can quickly arrive at the scene for processing. Through GPS positioning, the location of the intersection around the accident can be determined, the affected traffic area can be quickly identified, and emergency signal control can be performed to avoid secondary accidents. The minimum rectangle containing the location of the traffic accident is determined to minimize the impact on urban traffic, improve traffic efficiency, and identify the specific areas affected by the traffic accident for targeted traffic signal adjustments.

[0078] Furthermore, judging the road passability of the traffic accident and controlling the traffic signal based on the road passability means using the DeepLabv3+ algorithm to perform semantic segmentation on the road of the traffic accident after detecting the traffic accident, and identifying road elements including lane lines, sidewalk lines and green belts;

[0079] Traffic accidents are detected using the YOLOv4 algorithm, marked using detection boxes, and identified using lane lines as lanes occupied and lanes not occupied by traffic accidents. If there is an unoccupied lane, the traffic accident is judged to be passable; if there is no unoccupied lane, the traffic accident is judged to be impassable.

[0080] If the traffic accident makes the road impassable, the traffic signal in the marked direction of the intersection and the independent controller will be controlled to be a red light to stop vehicles from passing, and the red light in the driving direction will be controlled to flash continuously to remind vehicles that there is a traffic accident and they need to detour;

[0081] If the traffic accident is passable, the number of unoccupied lanes on the road is identified, and the proportion of unoccupied lanes to the total lanes on the road is calculated. The cumulative traffic arrival volume A(t) and the cumulative departure volume D(t) during peak hours are reduced proportionally. At the same time, the average delay d(t) and the queue length Q(t) are increased proportionally. The traffic signal cycle c and the green light time g are then substituted into the graph neural network for optimization and determined as the new traffic signal control strategy and implemented.

[0082] The DeepLabv3+ algorithm uses convolutional neural networks to classify images at the pixel level, and can accurately identify and segment various elements on the road. Through precise semantic segmentation, it can clearly identify the impact of traffic accidents on different parts of the road, providing basic data for subsequent judgments on road passability. The efficient target detection capability of the YOLOv4 algorithm can accurately identify the location and impact range of traffic accidents in real time. Combined with the semantic segmentation results, it can clearly determine whether the traffic accident occupies the lane and the number of occupied lanes, providing a basis for further judgment of road passability. By analyzing the semantic segmentation and target detection results, the occupation of the lane by the traffic accident can be determined, and whether the road can continue to be passed. It can quickly and accurately determine the impact of traffic accidents on road passability, reduce traffic interruption time caused by accidents, optimize traffic flow, adjust traffic signal control strategies according to the actual road conditions, ensure road traffic efficiency and safety, and dynamically adjust traffic signal control strategies to reduce the impact of traffic accidents on surrounding traffic and improve the flexibility and response speed of traffic management.

[0083] S3, the control center remotely monitors road data and traffic accidents, and responds and stores them in real time;

[0084] Specifically, the control center's remote monitoring of road data and traffic accidents means that the control center remotely receives road data and traffic accident analysis results collected by sensors and independent controllers through the network, and displays the road data in a visual chart in real time, and simultaneously manually analyzes the severity of the traffic accident based on the location of the traffic accident and the traffic accident image, and notifies relevant personnel to handle the traffic accident.

[0085] Sensors collect road traffic data in real time, including vehicle flow, speed, traffic accident information, etc., and transmit it to the control center through the network to ensure the timeliness and accuracy of the data, realize real-time monitoring and management of urban traffic conditions, and can quickly obtain and handle sudden traffic incidents, improve the response speed and efficiency of traffic management, and use data visualization technology to convert complex traffic data into intuitive charts, such as flow charts, heat maps, accident distribution maps, etc., which are convenient for managers to quickly understand and analyze. Through graphical display, managers can quickly grasp the current traffic conditions, discover potential problems, and make timely adjustments. Data visualization can also help discover long-term traffic trends and provide a basis for future traffic planning. On the basis of automated analysis, managers comprehensively judge the severity and urgency of the accident based on the accident scene images and location information, and decide on emergency response measures. Combining the advantages of artificial intelligence and manual analysis, the accuracy and effectiveness of traffic accident handling are improved. Through manual analysis, the shortcomings of the automated system in complex scenarios can be compensated to ensure the comprehensiveness and reliability of decision-making.

[0086] Furthermore, real-time response and storage means that after the control center receives the road data and traffic accident analysis results, they are stored in real time through the database, and the traffic signal control strategy formulated by the independent controller of each driving direction at the intersection is synchronously responded to, and the traffic signal control strategy is bound to the independent controller positioning and stored in a classified manner.

[0087] Real-time storage can ensure the integrity and timeliness of data, avoid data loss, and facilitate subsequent query and analysis. Through real-time storage, the control center can obtain the latest road and accident data at any time, provide accurate data support for traffic management, and improve decision-making efficiency. Synchronous response ensures that traffic signal control strategies can be executed in a timely manner, optimize traffic flow, and reduce the impact of traffic accidents on traffic flow. Through synchronous response, the control center can quickly adjust signal control strategies to ensure smooth traffic flow, reduce traffic congestion, and improve road traffic efficiency. Binding and classified storage can ensure that each signal control strategy can correspond to a specific independent controller for easy management and execution. Through binding and classified storage, the control center can clearly manage the signal control strategies of each intersection, ensure the accurate execution of the strategy, and improve the reliability and effectiveness of signal control.

[0088] based on Figure 2 Combined with the invention content of the above-mentioned road information collection data analysis method based on artificial intelligence, a complete process of independent controller for traffic signal optimization can be obtained, that is, collecting road data, calculating the cumulative traffic arrivals and cumulative departures in the driving direction, the average delay during peak hours and the queue length through an independent controller, and using the calculated parameters as graph neural network node features, optimizing the node features through the reinforcement learning Q-learnning algorithm, and setting the initial population according to the optimized node features and combining iterative optimization with the genetic algorithm, obtaining the population with the highest fitness through iterative optimization, and extracting traffic signal parameters as the optimal traffic signal control strategy.

[0089] Example 2

[0090] Reference Figure 3 , which is the second embodiment of the present invention, and which is different from the previous embodiment, provides a road information collection and data analysis system based on artificial intelligence, which includes:

[0091] Data collection module, used to deploy sensors to collect road data and perform pre-processing;

[0092] A traffic control module, which is used to optimize the traffic signal for each driving direction at the intersection based on the collected road data;

[0093] An accident detection module is used to detect traffic accidents on the road, analyze the passability of the traffic accidents, and control traffic signals based on the passability of the traffic accidents;

[0094] The monitoring and storage module is used to remotely receive road data and traffic accident analysis results, and store them in a classified manner through a database.

[0095] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0096] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0097] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0098] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A road information collection data analysis method based on artificial intelligence, characterized in that: include, Deploy sensors to collect road data and determine the direction of travel. Install independent controllers for each direction of travel to optimize traffic signals. Analyze traffic accidents based on collected road data, determine the road passability of the traffic accident, and control traffic signals based on the road passability; The control center remotely monitors road data and traffic accidents, and responds and stores them in real time; The deploying of sensors to collect road data, determining the road driving direction, and installing an independent controller for each road driving direction to optimize traffic signals refers to deploying sensors on the road to collect road data and pre-process the road data, and determining the driving direction according to the road data; Install independent controllers for traffic lights in each driving direction, and connect independent controllers at the same intersection through the network; Calculate the cumulative arrival volume A(t) and cumulative departure volume D(t) of peak-hour traffic in each driving direction: Where t is the cumulative time, For time The vehicle arrival rate, For time The vehicle departure rate; Calculate the average rush hour delay d(t) for each direction of travel: Where c is the traffic signal cycle, g is the green light time, x(t) is the saturated traffic flow at time t, q(t) is the vehicle arrival rate at time t, and S is the saturated vehicle arrival rate; Analyze the peak-hour queue length Q(t) for each driving direction based on road data: The independent controllers of each driving direction at the same intersection are used as nodes, and the influence between driving directions is used as edges to construct a graph neural network for the same intersection. If there is an influence between the driving directions between nodes, edge connections are constructed, and the cumulative traffic arrival volume A(t), cumulative departure volume D(t), average delay d(t) and queue length Q(t) during peak hours of each driving direction are calculated and recorded in the independent controller of each driving direction as node features; Extract the traffic flow in each driving direction and calculate the mean of the traffic flow between two connected nodes in the graph neural network as the weight of the edge connection; The graph neural network updates the hidden state of each node by aggregating adjacent nodes and edge connections, uses the back-propagation algorithm to train the graph neural network, defines the minimization loss function to optimize the node state, and uses the reinforcement learning Q-learnning algorithm to optimize the node features based on the optimized node state; The traffic signal cycle c, green light time g, vehicle arrival rate q(t) at time t, and saturated vehicle arrival rate S are taken as combinations, and all possible combinations are analyzed as the initial population according to the optimized node characteristics; Calculate the fitness of each combination in the initial population, select the combination with the highest fitness as the next generation parent combination, generate new combinations from the parent combinations through crossover and mutation and add them to the initial population; Iterate and repeat the selection, crossover and mutation operations until the preset number of iterations is reached or the fitness begins to converge, and then stop the iteration. The combination with the highest fitness in the population formed by stopping the iteration is taken as the optimal combination, and the traffic signal cycle c and green light time g in the optimal combination are extracted as the traffic signal control strategy for each driving direction; The analyzing of traffic accidents based on the collected road data includes: The YOLOv4 algorithm is used to detect vehicles, pedestrians and obstacles on the road in real time, and output a detection frame containing the object category and location; Define vehicles and pedestrians as traffic participants, and use convolutional neural networks to analyze traffic participant images to detect traffic accidents; If a traffic accident is detected, the location of the traffic accident is determined through GPS positioning, and emergency personnel are notified to handle the traffic accident. The location of the intersection around the traffic accident is simultaneously determined through GPS positioning, and the independent controllers of each driving direction in the intersection are connected as points to form a rectangle, and the minimum rectangle containing the location of the traffic accident is determined, and the independent controllers and driving directions in the minimum rectangle are marked; Determining the road passability of a traffic accident and controlling traffic signals based on the road passability means using the DeepLabv3+ algorithm to perform semantic segmentation on the traffic accident road after detecting the traffic accident, and identifying road elements including lane lines, sidewalk lines, and green belts; Traffic accidents are detected using the YOLOv4 algorithm, marked using detection boxes, and identified using lane lines as lanes occupied and lanes not occupied by traffic accidents. If there is an unoccupied lane, the traffic accident is judged to be passable; if there is no unoccupied lane, the traffic accident is judged to be impassable. If the traffic accident makes the road impassable, the traffic signal in the marked direction of the intersection and the independent controller will be controlled to be a red light to stop vehicles from passing, and the red light in the driving direction will be controlled to flash continuously to remind vehicles that there is a traffic accident and they need to detour; If the traffic accident is passable, the number of unoccupied lanes on the road is identified, and the proportion of unoccupied lanes to the total lanes on the road is calculated. The cumulative traffic arrival volume A(t) and the cumulative departure volume D(t) during peak hours are reduced proportionally. At the same time, the average delay d(t) and the queue length Q(t) are increased proportionally. The traffic signal cycle c and the green light time g are then substituted into the graph neural network for optimization and determined as the new traffic signal control strategy and implemented.

2. The method for analyzing road information collection data based on artificial intelligence according to claim 1, characterized in that: The sensors include lidar, camera, radar sensor, traffic flow sensor and GPS sensor.

3. The method for analyzing road information collection data based on artificial intelligence as claimed in claim 2, characterized in that: The control center remotely monitors road data and traffic accidents, which means that the control center remotely receives road data and traffic accident analysis results collected by sensors and independent controllers through the network, and displays the road data in a visual chart in real time, and simultaneously manually analyzes the severity of the traffic accident based on the location of the traffic accident and the traffic accident image, and notifies relevant personnel to handle the traffic accident.

4. The method for analyzing road information collection data based on artificial intelligence as claimed in claim 3, characterized in that: The real-time response and storage refers to the real-time storage of road data and traffic accident analysis results received by the control center through the database, synchronous response to the traffic signal control strategy formulated by the independent controller of each driving direction at the intersection, and binding the traffic signal control strategy with the positioning of the independent controller and storing them in a classified manner.

5. A road information collection data analysis system according to any one of claims 1 to 4, characterized in that: include, Data collection module, used to deploy sensors to collect road data and perform pre-processing; A traffic control module, which is used to optimize the traffic signal for each driving direction at the intersection based on the collected road data; An accident detection module is used to detect traffic accidents on the road, analyze the passability of the traffic accidents, and control traffic signals based on the passability of the traffic accidents; The monitoring and storage module is used to remotely receive road data and traffic accident analysis results, and store them in a classified manner through a database.

6. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that when the processor executes the computer program, the steps of the road information collection data analysis method based on artificial intelligence described in any one of claims 1 to 4 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the road information collection data analysis method based on artificial intelligence described in any one of claims 1 to 4 are implemented.

Images