Green wave optimization method based on cellular automaton under BIM-GIS-MLP

Abstract

The invention discloses a green wave optimization method based on a cellular automaton under BIM-GIS-MLP, and relates to the field of road digitization, and the method comprises the steps: segmenting a road according to a BIM and GIS platform; the terminal point information of the multiple journeys of each vehicle is counted based on the IOT, and corresponding traffic light calibration information and access information of the subsequent vehicles through the IOT are fed back to the subsequent vehicles; the BIM and GIS platform analyzes the average speed range passing through the multiple sections of continuous travel and calculates the sum of the duration of the green wave section; and calculating the optimization problem of the sum of the duration of the green wave section through linear programming, and solving the maximum value of the target function. According to the method, data networks of the Internet of Vehicles and the Internet of Things are accessed to realize data fusion and intercommunication, an MLP model is accessed to analyze individuals, and then the maximum value of the target function about the green wave road section time is obtained through linear programming, so that the traffic smoothness is improved, the energy loss and exhaust emission are reduced, and the urban environment quality is improved.

Description

Technical Field

[0001] The present invention relates to the field of highway digitization, and in particular to a green wave optimization method based on cellular automata under BIM-GIS-MLP. Background Art

[0002] During highway digitization, the measurement of green wave sections is affected by fluctuations in traffic flow. Failure to account for these fluctuations during measurement can lead to discrepancies between actual results and actual conditions. Furthermore, vehicle speed is affected by various factors, such as road conditions and weather conditions. Furthermore, improper signal timing is a significant factor. If signal timing is not adjusted promptly based on actual traffic conditions, the green wave effect can be poor, affecting measurement accuracy.

[0003] Nowadays, the timing of traffic lights is not in compliance with the normal operation regulations of the road. The inaccurate timing of traffic lights leads to measurement deviations in green wave sections, causing vehicles to drive out of the green wave sections. Summary of the Invention

[0004] The present invention provides a green wave optimization method based on cellular automata under BIM-GIS-MLP to solve the problems of the prior art.

[0005] In a first aspect, the present invention provides a green wave optimization method based on cellular automata under BIM-GIS-MLP, comprising: Divide the road into sections based on BIM and GIS platforms; The IoT collects the endpoint information of each vehicle's multiple trips and feeds back the corresponding traffic light calibration information to subsequent vehicles, which then access the information through the IoT. The BIM and GIS platform analyzes the average speed range of multiple consecutive trips and calculates the total duration of green wave sections; The optimization problem of the total duration of green wave sections is calculated through linear programming and the maximum value of the objective function is solved.

[0006] Furthermore, the road segmentation based on the BIM and GIS platform includes: Determine whether the current vehicle can pass the traffic lights of the subsequent journey. If it can, multiple traffic light roads that can be passed will be merged into one journey and included in the green wave section duration. If it cannot pass, one traffic light road in the current journey will be set as a journey and will not be included in the green wave section duration. The physical section boundaries will be divided according to the semantic information of the BIM components of bridges, tunnels, and intersections.

[0007] Furthermore, the current vehicle accesses the traffic light calibration information of the previous vehicle to determine whether it can still pass the traffic light of the subsequent trip at the current location. If it can pass, the assisted driving information is not modified. If it cannot pass, the speed range required to pass the merged trip is recalculated and a prompt is given to the driver.

[0008] Furthermore, the speed range required to pass the merged journey is recalculated and the driver is prompted, including that the required speed range includes different speed ranges under multiple segments, and it is determined whether the different speed ranges under each segment exceed the speed limit. If the speed limit is exceeded, the merged journey is divided into two or even multiple segments at the speeding section. If the speed limit is not exceeded, each segment with different speed ranges is merged into one journey.

[0009] Furthermore, the road segmentation based on the BIM and GIS platform includes: The road is divided into sections according to traffic lights, and the current speed and traffic flow data are used to determine whether to merge adjacent sections. When the speed and traffic flow data meet the merging requirements, the adjacent sections are merged into one trip, and the driver is given the corresponding speed prompt. When the speed and traffic flow data do not meet the merging requirements, the trip that does not meet the merging requirements will be segmented separately, and the adjacent sections that meet the merging requirements will be merged into one trip. The IoT traffic density heat map is integrated in real time, and automatic merging is triggered when the density difference between adjacent sections is less than the threshold.

[0010] Furthermore, for the two adjacent trips after the merger, the traffic light correction data of the previous vehicle is connected to determine whether the vehicle speed can pass the traffic light of the next trip within the calculated value range. When the speed is not speeding, it is further determined whether the traffic flow in the next trip is sufficient for the current vehicle to enter the next trip without waiting for the traffic light. When the speed is speeding, the current judgment condition is exited. When the traffic flow is sufficient, the two trips are merged into one trip, and the subsequent trips are judged cyclically, and the length of the green wave section is calculated based on the assisted driving speed prompt. When the traffic flow is not sufficient, the original trip segmentation is maintained.

[0011] Furthermore, the speed and navigation route of each vehicle are counted, and the traffic flow prediction is trained through the multi-layer perceptron in the server to predict the traffic flow in the subsequent time periods when driving to different travel locations, and the speed range of multiple continuous travel segments is given.

[0012] Furthermore, the multilayer perceptron is trained to predict traffic flow, specifically including the following steps: Access data collection information, including current vehicle speed, historical traffic flow and timestamp data; Data is cleaned and normalized, including: Using IoT data cleaning specifications to implement the following cleaning: For vehicle terminal data: eliminate abnormal sudden braking records with accelerations > 3m / s²; for roadside sensors: compensate for detection errors in rainy and foggy weather; A multi-layer perceptron model is constructed. The input layer selects multiple data type feature settings according to the node and time window, and the hidden layer uses an activation function for nonlinear transformation. The output layer is the speed range corresponding to the number of time windows. The MLP model is trained based on historical data. GIS road network topology encoding, including road segment IDs and upstream and downstream connections, is used as MLP input features. This is then combined with IoT vehicle speed time series data for joint training. Forward and backward propagation are iteratively performed to update model parameters until the loss function converges. Based on the current vehicle speed data, the trained MLP model is used to predict the speed range of multiple future trips. For multiple trips without waiting for traffic lights, the required speed range is set continuously; Output the speed range results of multiple trips of the current vehicle.

[0013] Furthermore, during the training of the MLP model, historical traffic flow data is screened out, and the traffic flow data and timestamp data of the corresponding trip from the target vehicle to the destination information section are selected. The MLP model is trained for the target vehicle, and multiple consecutive trips are set on a multi-core processor for independent training. After summarizing the results, the trained MLP model is output.

[0014] Furthermore, the optimization problem of calculating the total duration of green wave sections by linear programming and solving the maximum value of the objective function includes: While assisting in analyzing the speed range of future green wave sections for a single vehicle intervention in the MLP model, the cross-correlation impact between multiple vehicle speed ranges is analyzed, including the following steps: Set up a cellular automaton model to solve the cross-correlation effect and determine the model parameters of the cellular automaton model, including road length, number of lanes, vehicle density, maximum speed, and random deceleration probability; Initialize the distribution of vehicles on the road based on traffic observation data and call the data information to assign the initial speed to each vehicle; Setting the update rules of the cellular automaton model, including acceleration rules, deceleration rules, random deceleration rules, and lane change rules. It also includes extracting the number of lanes and slope curvature from the BIM model and writing them into the cellular automaton rule base, and calling the prompt speed data for each vehicle on the auxiliary green wave section. Dynamic graph generation: Real-time data lineage updates based on Apache Kafka.

[0015] # Cell velocity update rule (affected by BIM road slope) def update_speed(cell): max_speed = BIM.get_max_speed(cell.road_segment_id) # Get the design speed limit from BIM if cell.speed <max_speed: cell.speed += 1 * (1 - BIM.get_gradient(cell.position) / 10) #The greater the slope, the slower the acceleration The road is divided into GIS coordinate grids, with each cell carrying BIM lane attributes and IoT real-time traffic flow status. The vehicle status of all vehicles at each timestamp, including location and speed, is synchronously updated to obtain the latest simulated traffic flow dynamic evolution data; Obtain the traffic flow trend of the number of vehicles and speed over time; Load multiple vehicle speed data and vehicle data into the traffic flow change trend to obtain the predicted changes in traffic flow data in the future trip.

[0016] Furthermore, because the BIM and GIS platforms integrate the latest road construction information, after loading the data, they serve as a basic auxiliary platform, providing a strong data foundation for road optimization. The data information specifications of the BIM and GIS platforms are very helpful for road segmentation. Secondly, IoT records a variety of vehicle data information in multi-vehicle applications. The integration of IoT information facilitates the BIM and GIS platforms to assist vehicles in completing macro-driving optimization of road vehicles, providing a convenient traffic environment. In addition, BIM provides centimeter-level road geometry accuracy (such as curve radius and longitudinal slope), supporting the precise calculation of green wave speed ranges, and GIS realizes global spatial indexing, so that IoT vehicle trajectories are associated with road network topology in real time. IoT and cellular automation closed-loop optimization: vehicles upload the actual time they pass through traffic lights through V2I, dynamically calibrating the timing plan preset by BIM. The cellular model adjusts simulation parameters based on the real-time traffic flow of IoT, improving the feasibility of the linear programming solution.

[0017] The green wave optimization method based on cellular automata under BIM-GIS-MLP provided by the present invention is connected to the Internet of Vehicles and Internet of Things data networks to achieve data fusion and interoperability, and is connected to the MLP model to analyze individuals. Then, the maximum value of the objective function regarding the green wave section time is obtained through linear programming, thereby improving traffic fluency, reducing energy loss and exhaust emissions, and improving urban environmental quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the present invention, and do not constitute a limitation of the embodiments of the present invention. In the drawings: Figure 1 A flowchart of a green wave optimization method based on cellular automata under BIM-GIS-MLP is provided as an exemplary embodiment of the present invention.

[0019] Figure 2 A flowchart of the training process of the MLP model used in the green wave optimization method based on cellular automata under BIM-GIS-MLP is provided as an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0020] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent like or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention, as detailed in the appended claims.

[0021] First, the terms involved in the present invention are explained: A green wave section, also known as a green wave belt, refers to a traffic management measure on a specific road that, through the coordinated control of traffic lights, enables vehicles to pass through multiple green lights continuously within a certain speed range, reducing the number of stops made by vehicles during driving, thereby improving road efficiency and traffic smoothness.

[0022] In the existing technology, inaccurate measurement of traffic light countdown data on the driving section in the navigation map is the main reason for vehicles to drive out of the green wave section. For long-distance driving, the vehicle's green wave section judgment is limited to the adjacent section and often appears to be fleeting. Taking into account that highway construction projects do not fully utilize various data, the present invention helps vehicles enter the green wave section through the combined data network of the Internet of Vehicles and the Internet of Things, combined with auxiliary reminders.

[0023] The present invention uses third-party data to indirectly apply traffic data related to safety, calibrates the countdown display of traffic lights, and calculates the known remaining countdown segments in advance according to different settings of traffic lights. It assists drivers from an individual perspective to enter green wave sections with longer durations within the specified speed, while ensuring that a large number of road vehicles in batches enter the green wave sections for the longest total time, thereby improving traffic fluidity.

[0024] The specific application scenario of the present invention is auxiliary navigation of green wave sections of highway engineering.

[0025] The green wave optimization method based on cellular automaton under BIM-GIS-MLP provided by the present invention is intended to solve the above technical problems in the prior art.

[0026] The following describes in detail the technical solution of the present invention and how the technical solution of the present invention solves the above-mentioned technical problems using specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The following embodiments of the present invention are described in conjunction with the accompanying drawings.

[0027] Example 1: In the present invention, the data information of each platform is applied as follows: IoT end layer: IoT devices such as on-board OBD terminals, roadside units (RSUs), and geomagnetic sensors collect real-time vehicle dynamic data, including GPS location, instantaneous speed, acceleration, and navigation path; and road environment data, including signal light phases, queue lengths, and lane-level traffic flow. Data flow → transmitted to edge computing nodes via 5G / V2X for preprocessing.

[0028] BIM platform layer: Loads a detailed 3D model of the highway project, including lane geometry (curvature, slope, width), traffic facility attributes (signal coordinates, detector locations, speed limit signs), and data flow. The model is lightweighted and published to a GIS spatial database, where an attribute dependency graph is constructed: component ID → material batch → inspection report → maintenance record. GIS platform layer: Integrates BIM models with geospatial data to implement: road network topology analysis (path connectivity, road segmentation), dynamic data spatial matching (vehicle position mapping to lane-level accuracy), data flow → generation of traffic digital twins with spatial attributes, and dynamic selection of key fields based on graph node centrality; Among them, the design side: BIM model; the construction side: IoT sensor; the maintenance side: GIS spatiotemporal data; like Figure 1 As shown in the figure, from a macro perspective, the auxiliary optimization method for digitizing highways and entering green wave sections is as follows: Segment the roads based on BIM and GIS platforms, i.e. S1. Segment the roads in urban areas according to the traffic light settings; Based on IOT, the destination information of each vehicle's multiple journeys is counted, i.e., S2, the number of vehicles on each road segment is counted, and the destination of the vehicles on the road is analyzed based on the navigation data; The BIM and GIS platform analyzes the average speed range across multiple consecutive trips and calculates the total duration of the green wave sections, S3-S7, as follows: S3. Divide the trip into multiple segments for each vehicle, and feed the driving record data of the preceding vehicle in the same segment together with the preset countdown data of the traffic light to the navigation system; S4. The navigation system corrects the traffic light record data of the subsequent vehicle on the same trip; S5. Calculate the average speed range over multiple consecutive travel segments; S6. Analyze the traffic flow prediction data predicted by the multi-layer perceptron for each vehicle through multiple trips to obtain the duration of the green wave section in the discontinuous multiple trips; S7. Calculate the total green wave duration based on the sum of the number of vehicles that enter the next segment without waiting and pass through traffic lights; The optimization problem of the total duration of the green wave section is calculated through linear programming and the maximum value of the objective function is solved, that is, S8 and S9 are as follows: S8, cyclically calculating the total green wave duration of each next trip to establish an optimization problem; S9. Call linear programming to calculate the maximum value of the objective function of the optimization problem.

[0029] In the data verification phase: Load the metadata map, and missing fields will automatically trigger BIM model reverse query or on-site sensor supplementary sampling.

[0030] The following is an example of five-dimensional rule verification: Specifically, in step S6, the multilayer perceptron data is introduced to train the traffic flow prediction, such as Figure 2 As shown below: K1, access data collection information, including current vehicle speed, historical traffic flow and timestamp data; K2. Clean and normalize the data, including: Implementing the following cleaning methods using IoT data cleaning specifications: For vehicle terminal data, eliminate abnormal braking records with accelerations > 3m / s²; for roadside sensors, compensate for detection errors in rainy and foggy weather; K3. Construct a multi-layer perceptron model. The input layer selects multiple data type feature settings according to the node and time window. The hidden layer uses the activation function to perform nonlinear transformation. The output layer is the speed range corresponding to the number of time windows. K4. Train the MLP model based on historical data. GIS road network topology encoding, including road segment IDs and upstream and downstream connection relationships, is used as MLP input features. This is then combined with IoT vehicle speed time series data for joint training. Forward and backward propagation are iteratively performed to update model parameters until the loss function converges. K5: Use the trained MLP model to predict the speed range of multiple future trips based on the current speed data; K6. For multiple trips that do not require waiting for traffic lights, continuously set the required speed range; K7. Output the speed range results of the current vehicle's multiple travel segments.

[0031] Specific examples of steps K1-K7 are as follows: % Load data data = load('traffic_data.mat'); % Assume that the data is stored in the traffic_data.mat file X = data.X; % Input features: historical vehicle speed, traffic flow, traffic light timing, etc. y = data.y; % Output target: future speed range % Data preprocessing % Normalization X = (X - min(X)) / (max(X) - min(X)); y = (y - min(y)) / (max(y) - min(y)); % Divide the data set rng(1); % Set the random seed to ensure repeatable results idx = randperm(size(X, 1)); trainRatio = 0.7; validateRatio = 0.2; testRatio = 0.1; trainIdx = idx(1:floor(trainRatio * end)); validateIdx = idx(floor(trainRatio * end) + 1:floor((trainRatio +validateRatio) * end)); testIdx = idx(floor((trainRatio + validateRatio) * end) + 1:end); XTrain = X(trainIdx,:); yTrain = y(trainIdx, :); XValidate = X(validateIdx, :); yValidate = y(validateIdx, :); XTest = X(testIdx, :); yTest = y(testIdx, :); % Build MLP model hiddenLayerSize = 10; % Number of hidden layer neurons net = feedforwardnet(hiddenLayerSize); % Set training parameters net.trainFcn = 'trainlm'; % Levenberg-Marquardt optimization algorithm net.trainParam.epochs = 1000; % Maximum number of training rounds net.trainParam.goal = 1e-5; % target mean square error net.trainParam.max_fail = 10; % Maximum number of failures % Training model [net, tr]= train(net, XTrain', yTrain'); % Evaluate model % Validation set evaluation yValidatePred = net(XValidate'); validateMSE = mean((yValidatePred - yValidate') .^ 2); % Test set evaluation yTestPred = net(XTest'); testMSE = mean((yTestPred - yTest') .^ 2); % Predict future vehicle speed range % Assume we have new input data newX newX = ... % New input data, which must be in the same format as the training data newX = (newX - min(X)) / (max(X) - min(X)); % Normalization yPred = net(newX'); % Denormalized prediction results yPred = yPred * (max(y) - min(y)) + min(y); % Output prediction results disp('Predicted speed range:'); disp(yPred); A comparison chart of the predicted and actual values ​​obtained using MATLAB shows high consistency between the two. A linear programming optimization problem was solved for multiple vehicles, obtaining the total number of green wave sections with the longest duration. This ensures low vehicle energy consumption and reduces exhaust emissions. In this embodiment, the edge node processes IoT data with a response time of <50ms, meeting the real-time green wave decision-making requirements. The BIM model is converted to the CGCS2000 coordinate system and aligned with the GIS basemap.

[0032] The modules described as separate components may or may not be physically separate, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules may be selected to achieve the purpose of the present embodiment according to actual needs.

[0033] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0034] The above are merely embodiments of the present invention and are not intended to limit the present invention. It will be apparent to those skilled in the art that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are intended to be included within the scope of the claims of the present invention.

[0035] Other embodiments of the present invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow from the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the invention being indicated by the following claims.

[0036] It should be understood that the present invention is not limited to the exact construction described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof, which is limited only by the appended claims.

Claims

1. The green wave optimization method based on cellular automata under BIM-GIS-MLP is characterized by: include: Loading highway digital information into BIM and GIS platforms based on IOT sensors; Divide the road into sections based on BIM and GIS platforms; The IoT collects the endpoint information of each vehicle's multiple trips and feeds back the corresponding traffic light calibration information to subsequent vehicles, which then access the information through the IoT. The BIM and GIS platform analyzes the average speed range of multiple consecutive trips and calculates the total duration of green wave sections; The optimization problem of the total duration of green wave sections is calculated through linear programming and the maximum value of the objective function is solved.

2. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 1 is characterized in that: The road segmentation based on BIM and GIS platform includes: Determine whether the current vehicle can pass the traffic lights of the subsequent journey. If it can, multiple traffic light roads that can be passed will be merged into one journey and included in the green wave section duration. If it cannot pass, one traffic light road in the current journey will be set as a journey and will not be included in the green wave section duration. The physical section boundaries will be divided according to the semantic information of the BIM components of bridges, tunnels, and intersections.

3. The green wave optimization method based on cellular automaton under BIM-GIS-MLP according to claim 2 is characterized in that: The current vehicle accesses the traffic light calibration information of the previous vehicle to determine whether it can still pass the traffic light of the subsequent trip at the current location. If it can pass, the assisted driving information is not modified. If it cannot pass, the speed range required to pass the merged trip is recalculated and a prompt is given to the driver.

4. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 3 is characterized in that: The recalculation of the speed range required for passing the merged segment and the prompting of the driver include: the required speed range includes different speed ranges under multiple segments, and determining whether the different speed ranges under each segment exceed the speed limit; if exceeding the speed limit, the merged segment is divided into two or even multiple segments at the speed exceeding section; if not exceeding the speed limit, each segment of the different speed ranges is merged into one segment.

5. The green wave optimization method based on cellular automaton under BIM-GIS-MLP according to claim 1 is characterized in that: The road segmentation based on BIM and GIS platform includes: The road is divided into sections according to traffic lights, and the current speed and traffic flow data are used to determine whether to merge adjacent sections. When the speed and traffic flow data meet the merging requirements, the adjacent sections are merged into one trip, and the driver is given the corresponding speed prompt. When the speed and traffic flow data do not meet the merging requirements, the trip that does not meet the merging requirements will be segmented separately, and the adjacent sections that meet the merging requirements will be merged into one trip. The IoT traffic density heat map is integrated in real time, and automatic merging is triggered when the density difference between adjacent sections is less than the threshold.

6. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 5 is characterized in that: For the two adjacent trips after the merger, the traffic light correction data of the previous vehicle is connected to determine whether the vehicle speed can pass the traffic light of the next trip within the calculated value range. When the speed is not speeding, it is further determined whether the traffic flow in the next trip is sufficient for the current vehicle to enter the next trip without waiting for the traffic light. When the speed is speeding, the current judgment condition is exited. When the traffic flow is sufficient, the two trips are merged into one trip, and the subsequent trips are judged cyclically, and the length of the green wave section is calculated based on the assisted driving speed prompt. When the traffic flow is not sufficient, the original trip segmentation is maintained.

7. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 6 is characterized in that: For each vehicle, the vehicle speed and navigation route are counted, and the traffic flow prediction is trained through the multi-layer perceptron in the server to predict the traffic flow in the subsequent time periods when driving to different travel locations, and the speed range of multiple continuous travel segments is given.

8. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 7 is characterized in that: The multilayer perceptron is trained to predict traffic flow, specifically including the following steps: Access data collection information, including current vehicle speed, historical traffic flow and timestamp data; Data is cleaned and normalized, including: Using IoT data cleaning specifications to implement the following cleaning: For vehicle terminal data: eliminate abnormal sudden braking records with accelerations > 3m / s²; for roadside sensors: compensate for detection errors in rainy and foggy weather; A multi-layer perceptron model is constructed. The input layer selects multiple data type feature settings according to the node and time window, and the hidden layer uses an activation function for nonlinear transformation. The output layer is the speed range corresponding to the number of time windows. The MLP model is trained based on historical data. GIS road network topology encoding, including road segment IDs and upstream and downstream connections, is used as MLP input features. This is then combined with IoT vehicle speed time series data for joint training. Forward and backward propagation are iteratively performed to update model parameters until the loss function converges. Based on the current vehicle speed data, the trained MLP model is used to predict the speed range of multiple future trips. For multiple trips without waiting for traffic lights, the required speed range is set continuously; Output the speed range results of multiple trips of the current vehicle.

9. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 8 is characterized in that: During the training of the MLP model, historical traffic flow data is filtered out, and the traffic flow data and timestamp data of the corresponding trip from the target vehicle to the destination information section are selected. The MLP model is trained for the target vehicle, and multiple consecutive trips are set up on a multi-core processor for independent training. After summarizing the results, the trained MLP model is output.

10. The green wave optimization method based on cellular automata under BIM-GIS-MLP according to claim 9 is characterized in that: The optimization problem of calculating the total duration of green wave sections by linear programming and solving the maximum value of the objective function includes: While assisting in analyzing the speed range of future green wave sections for a single vehicle intervention in the MLP model, the cross-correlation impact between multiple vehicle speed ranges is analyzed, including the following steps: Set up a cellular automaton model to solve the cross-correlation effect and determine the model parameters of the cellular automaton model, including road length, number of lanes, vehicle density, maximum speed, and random deceleration probability; Initialize the distribution of vehicles on the road based on traffic observation data and call the data information to assign the initial speed to each vehicle; Setting the update rules of the cellular automaton model, including acceleration rules, deceleration rules, random deceleration rules, and lane change rules. It also includes extracting the number of lanes and slope curvature from the BIM model and writing them into the cellular automaton rule base, and calling the prompt speed data for each vehicle on the auxiliary green wave section. The road is divided into GIS coordinate grids, with each cell carrying BIM lane attributes and IoT real-time traffic flow status. The vehicle status of all vehicles at each timestamp, including location and speed, is synchronously updated to obtain the latest simulated traffic flow dynamic evolution data; Obtain the traffic flow trend of the number of vehicles and speed over time; Load multiple vehicle speed data and vehicle data into the traffic flow change trend to obtain the predicted changes in traffic flow data in the future trip.

Images