Urban traffic path planning method and system

Abstract

The invention discloses an urban traffic path planning method and system, and belongs to the field of urban traffic planning. The method solves the problem that the complexity and dynamic nature of an urban traffic system cannot be effectively handled in the prior art, can identify traffic modes and congestion hotspots through deep mining and analysis of a big data set, provides comprehensive data support for path planning, predicts future traffic demands by using a multi-dimensional dynamic model, and improves the prediction efficiency. A scientific basis is provided for path planning, an optimal path planning scheme can be generated by applying a path planning algorithm based on a prediction result of a multi-dimensional dynamic model, and a real-time driving path is provided for a user based on the obtained path planning scheme, so that the path planning of the user in the driving process is improved, and the user experience is improved. And a user is helped to better cope with a complex traffic environment.

Description

Technical Field

[0001] The present invention relates to the technical field of urban traffic planning, and specifically to an urban traffic path planning method and system. Background Art

[0002] Traditional urban traffic path planning methods are mainly based on static data and simple prediction models, and cannot effectively cope with the complexity and dynamics of the urban traffic system. For example, traditional methods usually only consider the topological structure of the road network and the current traffic flow, and it is difficult to accurately predict the changing trend of future traffic demand, nor can they comprehensively consider the impact of various traffic factors on path planning; therefore, they do not meet the existing needs, and for this reason, we propose an urban traffic path planning method and system. Summary of the Invention

[0003] The purpose of the present invention is to provide an urban traffic path planning method and system. By deeply mining and analyzing a large dataset, traffic patterns, congestion hotspots, and population flow rules are identified, and then through a multi-dimensional dynamic model, future traffic demand is accurately predicted, and based on this, an optimal path planning scheme is generated to improve the path planning of users during driving, and solve the problems raised in the above background art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions: An urban traffic path planning method, including:

[0005] Obtain a large dataset of the urban traffic system and conduct in-depth mining and analysis to obtain traffic parameter data, where the traffic parameter data includes traffic patterns, congestion hotspots, and population flow rules;

[0006] Use a multi-dimensional dynamic model to predict future traffic demand;

[0007] Based on the prediction results of the multi-dimensional dynamic model, combine the traffic parameter data, and apply a path planning algorithm to generate an optimal path planning scheme;

[0008] Based on the optimal path planning scheme, improve and adjust the path planning of users during driving.

[0009] Further, obtaining a large dataset of the urban traffic system includes the following steps:

[0010] Integrate the obtained large dataset of the urban traffic system, including traditional data and meteorological data, where the traditional data includes road network, traffic facilities, traffic flow, and population distribution;

[0011] Process the obtained large dataset, including cleaning, denoising, format unification, and data fusion;

[0012] After processing, deep learning technology is used to deeply mine and analyze the large dataset to identify traffic patterns, congestion hotspots, and crowd flow patterns, specifically as follows:

[0013] Construct a deep learning network model based on a convolutional neural network;

[0014] Adjust the parameters of the constructed deep learning network model, optimize the loss function, and perform deep learning network model training;

[0015] Based on the processed large dataset, identify traffic patterns, congestion hotspots, and crowd flow patterns through the trained deep learning network model.

[0016] Furthermore, process the obtained large dataset, specifically as follows:

[0017] Data cleaning, which is used to process errors, missing values, and outliers in the original large dataset, including:

[0018] Remove missing values: Process missing data through interpolation or mean filling;

[0019] Process outliers: Detect and remove outliers in the large dataset using statistical methods;

[0020] Data denoising, which is used to remove the noise components in the large dataset using data smoothing or filtering techniques;

[0021] Format unification, which is used to convert data from different sources and different formats into a unified format, including:

[0022] Data naming specification: Establish a unified data naming rule;

[0023] Data format conversion: Convert large datasets from different sources into a unified format;

[0024] Data fusion, which is used to integrate data from multiple data sources, including:

[0025] Data association and matching: Connect relevant data in different data sources through data association and matching techniques.

[0026] Furthermore, use a multi-dimensional dynamic model to predict future traffic demand, including the following steps:

[0027] Standardize the traffic parameter data obtained from deep mining and analysis to ensure that the input range of the traffic parameter data in the multi-dimensional dynamic model is consistent;

[0028] The multi-dimensional dynamic model combines time series analysis and neural network models to predict future traffic demand using the traffic parameter data.

[0029] Take the time distribution characteristics in the traffic parameter data as time series data and input them into time series analysis to analyze the periodic, trend, and seasonal characteristics of traffic demand;

[0030] Input the spatial characteristics in the traffic parameter data into the neural network model to capture the spatial distribution law and dynamic changes of traffic demand;

[0031] The multi-dimensional dynamic model integrates the prediction results of time series analysis and the neural network model through weighted average or stacking methods, and finally outputs the prediction results of future traffic demand;

[0032] Among them, the finally output prediction results include traffic demand volume, traffic flow, congestion degree, and travel time.

[0033] Furthermore, constructing a multi-dimensional dynamic model includes the following steps:

[0034] Combine time series analysis and neural network model to construct a multi-dimensional dynamic model;

[0035] Collect historical data: Collect historical data sets of historical traffic flow, road network, population distribution, and weather, and perform data cleaning and standardization processing;

[0036] Divide the historical data set into training set, validation set, and test set for model training and validation, specifically:

[0037] Time series analysis: Select appropriate time series models according to the characteristics of the large data set, including ARIMA and SARIMA, and use the historical data set to estimate the parameters of the time series model;

[0038] Diagnose and optimize the time series model through residual analysis and information criterion methods to ensure that the time series model captures the trends, seasonality, and periodic characteristics of the time series data in the large data set;

[0039] Neural network model: Select appropriate neural network architectures, including recurrent neural networks, long short-term memory networks, convolutional neural networks, or combined models of the above networks;

[0040] Use the training set data to train the neural network model, adjust the model parameters to minimize the loss function, and use the validation set data to validate the neural network model to evaluate the generalization ability and prediction accuracy of the model;

[0041] Model integration: The multi-dimensional dynamic model uses weighted average and stacking methods to integrate the prediction results of time series analysis and the neural network model, and outputs the prediction of future traffic demand by the multi-dimensional dynamic model.

[0042] Furthermore, generating an optimal path planning scheme includes the following steps:

[0043] Select a path planning algorithm, including the dynamic window method or the genetic algorithm;

[0044] Combine the prediction results of the multi-dimensional dynamic model, and comprehensively consider multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety;

[0045] Generate an optimal path planning scheme through the path planning algorithm;

[0046] Among them, the dynamic window method: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model, the trajectory generation and evaluation of the dynamic window are updated in real time. In the evaluation stage, path length, travel time, energy consumption, congestion level, and safety are used as evaluation indicators, and finally the optimal path is selected;

[0047] Genetic algorithm: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model, and comprehensively considering multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety, a large number of paths are evaluated and screened, and finally an optimal path is found.

[0048] Furthermore, based on the optimal path planning scheme, improve and adjust the user's path planning during driving, including the following steps:

[0049] Improve the user's path planning during driving based on the generated optimal path planning scheme;

[0050] At the same time, deploy a real-time traffic monitoring system. The traffic monitoring system uses sensors and camera devices to collect traffic data in real time;

[0051] Analyze the traffic data collected in real time and evaluate the implementation effect of the path planning scheme;

[0052] According to the evaluation results of the traffic data, adjust and optimize the path planning scheme in a timely manner.

[0053] Furthermore, applying the path planning algorithm to generate an optimal path planning scheme also includes:

[0054] S1: Obtain the data of the urban road network, draw a node map of the urban road network, set each intersection in the city as a node, set the starting point of departure as the current node; set the node closest to the destination as the target node;

[0055] S2: Obtain the path information between the current node and the target node at the departure moment, determine the initial planned path, and obtain the next node adjacent to the current node based on the initial planned path as the node to be evaluated;

[0056] S3: Obtain all other nodes adjacent to the node to be evaluated except the initial planned path from the road network node map as alternative nodes. The node next to the neighbor of the node to be evaluated on the initial planned path is the node within the plan;

[0057] S4: Update the traffic flow, weather conditions, accidents, and road maintenance information obtained in real time during driving, calculate the passing costs between the node to be evaluated and each alternative node, and select the route between the alternative node with the minimum passing cost among all alternative nodes and the node to be evaluated as the alternative planned sub-path;

[0058] S5: When the driving distance of the vehicle from the node to be evaluated during driving on the initial planned path is less than 1 km, obtain the latest traffic flow, accident information, and road condition maintenance information between the node to be evaluated and the node within the plan, and conduct path evaluation. When it is judged whether the obtained path evaluation result is greater than the preset path evaluation result;

[0059] If so, it is determined that the initial planned path is still the optimal path. At the same time, control the vehicle to drive towards the path between the nodes within the plan after passing the node to be evaluated, and set the node within the plan as the new node to be evaluated, and repeat the steps of S3 and S4;

[0060] If not, re-evaluate and compare the initial planned path and the alternative planned sub-path, select the one with the better path evaluation result between the two as the actual driving path, and update the initial planned path based on the actual driving path;

[0061] S6: Repeat the above steps until the node to be evaluated coincides with the target node, and stop path planning.

[0062] Furthermore, the passing costs between the node to be evaluated and each alternative node include the path passing costs between the node to be evaluated and each alternative node, and the waiting time costs generated by passing through the traffic lights where the alternative nodes are located;

[0063] The following formula is used to calculate the waiting time costs generated by passing through the traffic lights where each alternative node is located:

[0064]

[0065] Among them, S is the path passing cost, v i is the estimated speed of the vehicle within the current time period, calculated by averaging the running speeds of all vehicles in the traffic area continuously for seven days at the same time period, t i is the green light time during which the vehicle passes through the intersection in one direction at the intersection, and T is the total sum of the green light and yellow light times for different directions of the same intersection, What is calculated is at time ti The probability of arriving at and passing through an intersection within a certain time Calculated to be at time t i The probability of not being able to pass through the intersection within a certain time; What is calculated is the waiting time at the intersection.

[0066] An urban traffic path planning system, applied to an urban traffic path planning method, the system includes a data integration unit, a path planning unit, and a path application unit;

[0067] The data integration unit is configured to integrate the acquired large dataset and perform processing, in-depth mining, and analysis;

[0068] The path planning unit is configured to use a multi-dimensional dynamic model to predict future traffic demands and generate an optimal path planning scheme according to the prediction results;

[0069] The path application unit is configured to combine the path planning scheme of the path planning unit to improve the path planning of the user during driving;

[0070] Among them, the path planning unit includes:

[0071] A model construction module configured to construct a multi-dimensional dynamic model by combining time series analysis and a neural network model;

[0072] A demand prediction module configured to input the large dataset that has been deeply mined and analyzed by the data integration unit into the multi-dimensional dynamic model. The multi-dimensional dynamic model uses the methods of weighted average and stacking to integrate the prediction results of the time series analysis and the neural network model, and outputs the prediction of the multi-dimensional dynamic model for future traffic demands;

[0073] A scheme generation module configured to generate an optimal path planning scheme according to the path planning algorithm and in combination with the prediction results of the demand prediction module;

[0074] The path application unit includes:

[0075] A path optimization module configured to improve the path planning of the user during driving based on the generated path planning scheme;

[0076] A monitoring and feedback module configured to collect traffic data in real time through the deployed real-time traffic monitoring system and perform analysis, evaluate the implementation effect of the path planning scheme through the analysis, and feedback the evaluation results to the staff at the terminal. The staff adjusts and optimizes the path planning scheme in a timely manner according to the evaluation results.

[0077] Compared with the prior art, the beneficial effects of the present invention are:

[0078] By obtaining a large dataset of the urban transportation system and conducting in-depth mining and analysis, the present invention can comprehensively and accurately grasp the current compliance rate of urban transportation. Using a multi-dimensional dynamic model, it can predict future traffic demands. Based on the prediction results of the multi-dimensional dynamic model, applying a path planning algorithm can generate an optimal path planning scheme, which takes into account the changes in future traffic conditions and provides a more efficient driving route for users. And based on the obtained path planning scheme, it can provide real-time path guidance for users, helping users better cope with complex traffic environments and enhancing the driving experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 is a schematic flowchart of the urban traffic path planning method of the present invention;

[0080] Figure 2 is a schematic diagram of the overall structure of the urban traffic path planning system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0081] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0082] To solve the technical problems that the existing urban traffic path planning methods are mainly based on static data and simple prediction models, unable to effectively cope with the complexity and dynamics of the urban traffic system, difficult to accurately predict the changing trend of future traffic demands, and unable to comprehensively consider the influence of various traffic factors on path planning, please refer to Figure 1 - Figure 2 , the present embodiment provides the following technical solutions:

[0083] An urban traffic path planning method, the path planning method generates a path planning scheme based on the prediction results of a multi-dimensional dynamic model for future traffic demands, and provides path planning for a user driving a vehicle;

[0084] The path planning method includes the following steps:

[0085] Obtain a large dataset of the urban traffic system and conduct in-depth mining and analysis to obtain traffic parameter data, and the traffic parameter data includes traffic patterns, congestion hotspots, and population flow rules;

[0086] Use a multi-dimensional dynamic model to predict future traffic demands;

[0087] Based on the prediction results of the multi-dimensional dynamic model, combine the traffic parameter data, and apply a path planning algorithm to generate an optimal path planning scheme;

[0088] Improve and adjust the path planning of users during driving based on an optimized path planning scheme.

[0089] The technical effects of the above are as follows: By deeply mining and analyzing the large dataset of the urban traffic system obtained, traffic patterns and congestion hotspots can be identified, thus providing comprehensive data support for path planning. Combining time series analysis and neural network models can accurately predict the changing trends of future traffic demands. Based on this, an optimized path planning scheme is generated, which can provide real-time path guidance for users, thereby improving the path planning of users during driving.

[0090] Obtain the large dataset of the urban traffic system, including the following steps:

[0091] Integrate the large dataset of the urban traffic system obtained, including traditional data and meteorological data, where the traditional data includes road networks, traffic facilities, traffic flows, and population distributions;

[0092] Process the obtained large dataset, including cleaning, denoising, format unification, and data fusion, specifically:

[0093] Data cleaning is used to process errors, missing values, and outliers in the original large dataset, including:

[0094] Remove missing values: Process the missing data through interpolation or mean filling to ensure the integrity of the data;

[0095] Process outliers: Use statistical methods to detect and remove outliers in the large dataset, improve the data quality, and prevent adverse effects on subsequent analysis;

[0096] Data denoising is used to remove the noise components in the large dataset using data smoothing or filtering techniques, that is, those redundant information that is irrelevant to the true characteristics of the data or interferes with the true characteristics of the data, so as to enhance the accuracy and reliability of the data;

[0097] Format unification is used to convert data from different sources and different formats into a unified format, including:

[0098] Data naming specification: Formulate a unified data naming rule to ensure that the data can be accurately traced back during subsequent use;

[0099] Data format conversion: Convert large datasets from different sources into a unified format to ensure the compatibility of the data in different platforms or script processing;

[0100] Data fusion is used to integrate data from multiple data sources, including:

[0101] Data Association and Matching: Through data association and matching technologies, relevant data from different data sources are connected to provide richer information;

[0102] After processing, deep learning technologies are used to deeply mine and analyze large datasets to identify traffic patterns, congestion hotspots, and crowd flow patterns, specifically:

[0103] Construct a deep learning network model based on a convolutional neural network;

[0104] Adjust the parameters of the constructed deep learning network model, optimize the loss function, and perform training on the deep learning network model. Among them, through parameter adjustment and loss function optimization, the training effect and generalization ability of the deep learning network model are improved to ensure that the model can accurately identify traffic patterns and crowd flow patterns in practical applications;

[0105] Based on the processed large dataset, identify traffic patterns, congestion hotspots, and crowd flow patterns through the trained deep learning network model.

[0106] The technical effects of the above content are as follows: Integrating traditional data (road network, traffic facilities, traffic flow, population distribution) and meteorological data, and through data cleaning, denoising, format unification, and fusion processing, the heterogeneity and quality problems of large datasets can be solved, providing a high-quality data foundation for subsequent analysis. Constructing a deep learning model based on a convolutional neural network (CNN) can efficiently process complex large datasets, thereby automatically extracting traffic patterns and crowd flow characteristics, accurately identifying congestion hotspot areas. Based on the above design, it provides a scientific basis for urban traffic route planning, can discover potential congestion points in advance, and provides support for formulating targeted optimization strategies.

[0107] Use a multi-dimensional dynamic model to predict future traffic demand, including the following steps:

[0108] Standardize the traffic parameter data obtained from deep mining and analysis to ensure that the input range of traffic parameter data in the multi-dimensional dynamic model is consistent;

[0109] The multi-dimensional dynamic model combines time series analysis and neural network models to predict future traffic demand using traffic parameter data;

[0110] Use the time distribution characteristics (such as traffic patterns) in the traffic parameter data as time series data and input them into time series analysis to analyze the periodic, trend, and seasonal characteristics of traffic demand;

[0111] Input the spatial characteristics (such as congestion hotspots and crowd flow patterns) in the traffic parameter data into the neural network model to capture the spatial distribution law and dynamic changes of traffic demand;

[0112] The multi-dimensional dynamic model integrates the prediction results of time series analysis and neural network models through weighted averaging or stacking methods, and finally outputs the prediction results of future traffic demand, including traffic demand volume, traffic flow, congestion level, and travel time;

[0113] Among them, constructing the multi-dimensional dynamic model includes the following steps:

[0114] Combine time series analysis and neural network models to construct a multi-dimensional dynamic model;

[0115] Collect historical data: Collect historical data sets of historical traffic flow, road network, population distribution, and weather, and perform data cleaning and standardization processing;

[0116] Divide the historical data set into a training set, a validation set, and a test set for model training and validation. Specifically:

[0117] Time series analysis: Select appropriate time series models according to the characteristics of the large data set, including ARIMA and SARIMA, and use the historical data set to estimate the parameters of the time series model;

[0118] Diagnose and optimize the time series model through residual analysis and information criterion methods to ensure that the time series model captures the trends, seasonality, and periodic characteristics of the time series data in the large data set;

[0119] Neural network model: Select appropriate neural network architectures, including recurrent neural networks, long short-term memory networks, convolutional neural networks, or combined models of the above networks;

[0120] Use the training set data to train the neural network model, adjust the model parameters to minimize the loss function, and use the validation set data to validate the neural network model to evaluate the generalization ability and prediction accuracy of the model;

[0121] Model integration: The multi-dimensional dynamic model uses weighted averaging and stacking methods to integrate the prediction results of time series analysis and neural network models, and outputs the prediction of future traffic demand by the multi-dimensional dynamic model.

[0122] The technical effects of the above content are as follows: By collecting and processing historical data sets, a rich data foundation is provided for model training and validation, ensuring that the model can capture the long-term trends and periodic changes in traffic demand. Appropriate time series models (such as ARIMA and SARIMA) are selected, and parameter estimation and model diagnosis are carried out to ensure that the model can accurately describe the time series characteristics of traffic demand, including trends, seasonality, and periodicity. Appropriate neural network architectures (such as recurrent neural networks, long short-term memory networks, convolutional neural networks) are selected, and the model parameters are adjusted through training to minimize the loss function. Moreover, the generalization ability and prediction accuracy of the model are evaluated. Combining time series analysis and neural network models, a multi-dimensional dynamic model is constructed, which can comprehensively consider various influencing factors of traffic demand, improve the accuracy and reliability of prediction. By adopting the methods of weighted average and stacking, the prediction results of time series analysis and neural network models are integrated, and the prediction of future traffic demand by the multi-dimensional dynamic model is output, improving the stability and accuracy of prediction, providing a scientific basis for the path planning of users during driving, being able to detect potential changes in traffic demand in advance, and providing support for formulating targeted optimization strategies.

[0123] Generate an optimal path planning scheme, including the following steps:

[0124] Select a path planning algorithm, including the dynamic window method or the genetic algorithm. Among them, the genetic algorithm is used for global path optimization and can quickly screen out excellent path schemes, while the dynamic window method is used for local path adjustment to ensure the feasibility and safety of the path in a dynamic environment;

[0125] Combine the prediction results of the multi-dimensional dynamic model and comprehensively consider multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety;

[0126] Generate an optimal path planning scheme through the path planning algorithm;

[0127] Among them, the dynamic window method: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model, the trajectory generation and evaluation of the dynamic window are updated in real time. In the evaluation stage, path length, travel time, energy consumption, congestion level, and safety are used as evaluation indicators, and finally the optimal path is selected;

[0128] Genetic algorithm: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model and comprehensively considering multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety, a large number of paths are evaluated and screened, and finally an optimal path is found.

[0129] The technical effects of the above content are as follows: By selecting the dynamic window method or the genetic algorithm and combining with the prediction results of the multi-dimensional dynamic model, it can be dynamically adjusted according to real-time traffic demands and congestion conditions, adapting to the complex and changeable urban traffic environment. At the same time, by comprehensively considering multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety, a path planning scheme can be generated to meet the requirements of the shortest path or the fastest path.

[0130] Based on the optimal path planning scheme, improve and adjust the path planning of the user during driving, including the following steps:

[0131] Improve the path planning of the user during driving based on the generated optimal path planning scheme;

[0132] At the same time, deploy a real-time traffic monitoring system, and the traffic monitoring system uses sensors and camera devices to collect traffic data of vehicles in real time;

[0133] Analyze the traffic data collected in real time and evaluate the implementation effect of the path planning scheme;

[0134] According to the evaluation results of the traffic data, adjust and optimize the path planning scheme in a timely manner.

[0135] The technical effects of the above content are as follows: By collecting traffic data of vehicles through the real-time traffic monitoring system, it is possible to understand road condition information in real time, such as road congestion level, vehicle speed, etc. During driving, if the real-time traffic data changes, such as a previously unobstructed road becomes congested, the path planning scheme can be adjusted and optimized in a timely manner to provide a new optimal path for the user, further enhancing the driving experience.

[0136] Furthermore, applying the path planning algorithm to generate the optimal path planning scheme also includes:

[0137] S1: Obtain data of the urban road network, draw a node map of the urban road network, set each intersection in the city as a node, set the starting point of departure as the current node; set a node closest to the destination as the target node;

[0138] S2: Obtain the path information between the current node and the target node at the departure moment, determine the initial planned path, and obtain the next node adjacent to the current node based on the initial planned path as the node to be evaluated;

[0139] S3: Obtain all other nodes adjacent to the node to be evaluated except the initial planned path from the road network node map as alternative nodes, and the next node adjacent to the node to be evaluated on the initial planned path is the node within the plan;

[0140] S4: Update the traffic flow, weather conditions, accidents, and road maintenance information obtained in real time during driving, calculate the travel costs between the node to be evaluated and each alternative node, and select the route between the alternative node with the minimum travel cost among all alternative nodes and the node to be evaluated as the alternative planned sub-path;

[0141] S5: When the driving distance of the vehicle from the node to be evaluated is less than 1 km during the driving process on the initial planned path, obtain the latest traffic flow, accident information, and road condition maintenance information between the node to be evaluated and the nodes within the plan, and conduct path evaluation. Determine whether the obtained path evaluation result is greater than the preset path evaluation result;

[0142] If so, determine that the initial planned path is still the optimal path. At the same time, control the vehicle to drive along the path between the nodes within the plan after passing the node to be evaluated, and set the nodes within the plan as the new nodes to be evaluated, repeating the steps of S3 and S4;

[0143] If not, re-evaluate and compare the initial planned path and the alternative planned sub-path, select the one with the better path evaluation result as the actual driving path, and update the initial planned path based on the actual driving path;

[0144] S6: Repeat the above steps until the node to be evaluated coincides with the target node, and stop the path planning.

[0145] The principles and effects of the above technical solution are as follows: When determining the initial planned path, usually only basic path information such as distance and road surface width is considered to determine the optimal path. However, during the driving process, traffic flow will change, and factors such as weather changes, traffic accidents, and road repairs will affect the initial planned path, which may no longer be the optimal path. Therefore, it is necessary to follow up and adjust in real time. Through the urban road network node map, the path between the starting point and the target node is divided according to intersections. After the vehicle starts from the starting point, the initial planned path is obtained, and the next node adjacent to the current node is used as the node to be evaluated. The alternative nodes are evaluated from the node to be evaluated to other adjacent nodes except the initial planned path, and alternative planned sub-paths are evaluated. This way of advance planning can ensure that during the period before reaching the node to be evaluated, if traffic jams, accidents, or other emergencies occur on the initial planned path, resulting in the path evaluation result of the initial planned path not meeting expectations, it is necessary to reconsider whether to continue along the initial planned path or adjust the planned path in a timely manner. At this time, the path conditions of the initial planned path and the alternative planned sub-paths are compared again, and a better path is selected as the route to be re-selected after reaching the node to be evaluated. This can quickly select a suitable new path when unexpected situations occur on the initial planned path, improve the smoothness of traffic operation, adapt to complex and changeable traffic demands and environmental changes, and ensure the real-time and effectiveness of the solution.

[0146] Furthermore, the passing cost from the node to be evaluated to each alternative node includes the path passing cost from the node to be evaluated to each alternative node and the waiting time cost generated by passing through the traffic lights where the alternative nodes are located.

[0147] The following formula is used to calculate the waiting time cost generated by passing through the traffic lights where each alternative node is located from the node to be evaluated:

[0148]

[0149] Among them, S is the path passing cost, and the path passing cost can be considered as the factors hindering the path passing process. v i is the estimated speed of the vehicle during the current period, which is calculated by the average running speed of all vehicles in the same time period within seven consecutive days in this traffic area. t i is the green light time when the vehicle passes through the intersection in one direction at the intersection, and T is the total sum of all green lights and yellow lights for different directions of the same intersection. What is calculated is the probability of arriving at and passing through the intersection within time t i time. What is calculated is within time t iThe probability of not being able to pass through the intersection within a certain time; What is calculated is the waiting time at the intersection.

[0150] The principle and effect of the above solution are as follows: When evaluating the traffic cost between the node to be evaluated and each alternative node, two aspects of factors need to be considered simultaneously. One is the path traffic cost between the node to be evaluated and each alternative node, which takes into account factors that affect normal driving such as congestion, weather, and road conditions during normal driving on the road section. On the other hand, it is the time and travel cost that need to be delayed during the process of waiting for traffic lights. And there is a probability problem in waiting for traffic lights. If you happen to encounter a green light, you don't need to wait. If you encounter a red light, you need to wait. Therefore, by calculating the waiting time cost generated by the traffic lights where the node to be evaluated and each alternative node are located, it can play a good guiding role in path planning, so as to flexibly adapt to traffic conditions, achieve real-time regulation, and improve urban road traffic.

[0151] Specifically, this embodiment also provides an urban traffic path planning system, which is applied to the urban traffic path planning method. The system includes a data integration unit, a path planning unit, and a path application unit;

[0152] The data integration unit is configured to integrate the acquired large data set and perform processing, in-depth mining, and analysis;

[0153] The path planning unit is configured to use a multi-dimensional dynamic model to predict future traffic demands and generate an optimal path planning scheme according to the prediction results;

[0154] The path application unit is configured to improve the path planning of users during driving in combination with the path planning scheme of the path planning unit;

[0155] Among them, the path planning unit includes:

[0156] The model construction module is configured to construct a multi-dimensional dynamic model by combining time series analysis and a neural network model;

[0157] The demand prediction module is configured to input the large data set that has been deeply mined and analyzed by the data integration unit into the multi-dimensional dynamic model. The multi-dimensional dynamic model uses the methods of weighted average and stacking to integrate the prediction results of time series analysis and the neural network model, and outputs the prediction of the multi-dimensional dynamic model for future traffic demands;

[0158] The scheme generation module is configured to generate an optimal path planning scheme according to the path planning algorithm and in combination with the prediction results of the demand prediction module;

[0159] The path application unit includes:

[0160] A path optimization module, configured to improve the path planning of a user during driving based on the generated path planning scheme;

[0161] A monitoring feedback module, configured to collect traffic data in real time through a deployed real-time traffic monitoring system, analyze the data, evaluate the implementation effect of the path planning scheme through the analysis, and feedback the evaluation result to the staff at the terminal. The staff can adjust and optimize the path planning scheme in a timely manner according to the evaluation result.

[0162] Working principle: By deeply mining and analyzing the data set, an accurate data basis is provided for path planning, making the generated path planning scheme more in line with the actual traffic conditions. By constructing a multi-dimensional dynamic model, future traffic demands can be predicted, and thus a path planning scheme can be generated according to the prediction results of the multi-dimensional dynamic model. Through the generated path planning scheme, the path planning of the user during driving can be improved, thereby helping the user better cope with complex traffic environments and enhancing the user's driving experience.

[0163] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0164] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention.

Claims

1. A method for urban traffic route planning, characterized in that, Including: Obtain a large dataset of the urban transportation system, conduct in-depth mining and analysis, and obtain traffic parameter data, where the traffic parameter data includes traffic patterns, congestion hotspots, and population flow patterns; Use a multi-dimensional dynamic model to predict future traffic demands; Based on the prediction results of the multi-dimensional dynamic model, combine with the traffic parameter data, and apply a path planning algorithm to generate an optimal path planning scheme; Based on the optimal path planning scheme, improve and adjust the path planning of users during driving.

2. The urban traffic route planning method according to claim 1, wherein: Obtain a large dataset of the urban transportation system, including the following steps: Integrate the obtained large dataset of the urban transportation system, including traditional data and meteorological data, where the traditional data includes road networks, traffic facilities, traffic flow, and population distribution; Process the obtained large dataset, including cleaning, denoising, format unification, and data fusion; After the processing is completed, use deep learning technology to conduct in-depth mining and analysis of the large dataset, and identify traffic patterns, congestion hotspots, and population flow patterns. Specifically: Construct a deep learning network model based on a convolutional neural network; Adjust the parameters of the constructed deep learning network model, optimize the loss function, and perform deep learning network model training; Based on the processed large dataset, identify traffic patterns, congestion hotspots, and population flow patterns through the trained deep learning network model.

3. The urban traffic path planning method according to claim 2, wherein: Process the obtained large dataset, specifically: Data cleaning, used to process errors, missing values, and outliers in the original large dataset, including: Remove missing values: Process missing data through interpolation or mean filling; Process outliers: Use statistical methods to detect and remove outliers in the large dataset; Data denoising, used to remove the noise components in the large dataset by using data smoothing or filtering techniques; Format unification, used to convert data from different sources and different formats into a unified format, including: Data naming specification: Develop a unified data naming rule; Data format conversion: Convert large datasets from different sources into a unified format; Data fusion, used to integrate data from multiple data sources, including: Data association and matching: Connect relevant data in different data sources through data association and matching techniques.

4. The urban traffic path planning method according to claim 1, wherein: Use a multi-dimensional dynamic model to predict future traffic demands, including the following steps: Standardize the traffic parameter data obtained from in-depth mining and analysis to ensure that the input range of the traffic parameter data in the multi-dimensional dynamic model is consistent; The multi-dimensional dynamic model combines time series analysis and neural network models, and uses the traffic parameter data to predict future traffic demands; Take the time distribution characteristics in the traffic parameter data as time series data and input them into the time series analysis to analyze the periodic, trend, and seasonal characteristics of traffic demands; Input the spatial characteristics in the traffic parameter data into the neural network model to capture the spatial distribution law and dynamic changes of traffic demands; The multi-dimensional dynamic model integrates the prediction results of the time series analysis and the neural network model through weighted average or stacking methods, and finally outputs the prediction results of future traffic demands; Among them, the finally output prediction results include traffic demand volume, traffic flow, congestion degree, and travel time.

5. The urban traffic path planning method according to claim 4, characterized in that: Construct a multi-dimensional dynamic model, including the following steps: Combine time series analysis and neural network models to construct a multi-dimensional dynamic model; Collect historical data: Collect historical datasets of historical traffic flow, road network, population distribution, and weather, and perform data cleaning and standardization processing; Divide the historical dataset into a training set, a validation set, and a test set for model training and validation. Specifically: Time series analysis: Select appropriate time series models according to the characteristics of the large dataset, including ARIMA and SARIMA, and use the historical dataset to estimate the parameters of the time series model; Diagnose and optimize the time series model through residual analysis and information criterion methods to ensure that the time series model captures the trends, seasonality, and periodic characteristics of the time series data in the large dataset; Neural network model: Select appropriate neural network architectures, including recurrent neural networks, long short-term memory networks, convolutional neural networks, or combined models of the above networks; Use the training set data to train the neural network model, adjust the model parameters to minimize the loss function, and use the validation set data to validate the neural network model to evaluate the generalization ability and prediction accuracy of the model; Model integration: The multi-dimensional dynamic model uses weighted average and stacking methods to integrate the prediction results of time series analysis and neural network models, and outputs the prediction of the multi-dimensional dynamic model for future traffic demand.

6. The urban traffic path planning method according to claim 1, characterized in that: Generate an optimal path planning scheme, including the following steps: Select a path planning algorithm, including the dynamic window method or genetic algorithm; Combine the prediction results of the multi-dimensional dynamic model and comprehensively consider multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety; Generate an optimal path planning scheme through the path planning algorithm; Among them, the dynamic window method: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model, the trajectory generation and evaluation of the dynamic window are updated in real time. In the evaluation stage, path length, travel time, energy consumption, congestion level, and safety are used as evaluation indicators, and finally the optimal path is selected; Genetic algorithm: In urban traffic path planning, according to the prediction results of the multi-dimensional dynamic model and comprehensively considering multi-objective factors such as path length, travel time, energy consumption, congestion level, and safety, a large number of paths are evaluated and screened, and finally an optimal path is found.

7. The urban traffic path planning method according to claim 1, characterized in that: Based on the optimal path planning scheme, improve and adjust the user's path planning during driving, including the following steps: Improve the user's path planning during driving based on the generated optimal path planning scheme; At the same time, deploy a real-time traffic monitoring system, and the traffic monitoring system uses sensors and camera devices to collect traffic data in real time; Analyze the traffic data collected in real time and evaluate the implementation effect of the path planning scheme; According to the evaluation results of the traffic data, adjust and optimize the path planning scheme in a timely manner.

8. The urban traffic path planning method according to claim 1, characterized in that: Applying the path planning algorithm to generate an optimal path planning scheme also includes: S1: Obtain the data of the urban road network, draw a node map of the urban road network, set each intersection in the city as a node, set the starting point of departure as the current node; set the node closest to the destination as the target node; S2: Obtain the path information between the current node and the target node at the departure time, determine the initial planned path, and obtain the next node adjacent to the current node based on the initial planned path as the node to be evaluated; S3: Obtain all other nodes adjacent to the node to be evaluated except the initial planned path from the road network node graph as alternative nodes, and the next node on the initial planned path that is a neighbor of the node to be evaluated is an in-planned node; S4: Update the traffic flow, weather conditions, accidents, and road maintenance information obtained in real time during driving, calculate the passing costs between the node to be evaluated and each alternative node, and select the route between the node to be evaluated and the alternative node with the minimum passing cost among all alternative nodes as the alternative planned sub-path; S5: When the driving distance of the vehicle from the node to be evaluated during driving on the initial planned path is less than 1 km, obtain the latest traffic flow, accident information, and road condition maintenance information from the node to be evaluated to the in-planned node, and perform path evaluation. When it is judged whether the obtained path evaluation result is greater than the preset path evaluation result; If so, determine that the initial planned path is still the optimal path. At the same time, control the vehicle to drive towards the path between the in-planned nodes after passing the node to be evaluated, and set the in-planned node as the new node to be evaluated, and repeat the steps of S3 and S4; If not, re-evaluate and compare the initial planned path and the alternative planned sub-path, select the one with the better path evaluation result between the two as the actual driving path, and update the initial planned path based on the actual driving path; S6: Repeat the above steps until the node to be evaluated coincides with the target node, and stop path planning.

9. The urban traffic path planning method according to claim 8, wherein: The passing costs between the node to be evaluated and each alternative node include the path passing costs between the node to be evaluated and each alternative node, and the waiting time costs generated by passing through the traffic lights where the alternative nodes are located; Use the following formula to calculate the waiting time costs generated by passing through the traffic lights where each alternative node is located from the node to be evaluated: Among them, S is the path passing cost, and v i is the estimated speed of the vehicle within the current time period, which is obtained by calculating the average running speed of all vehicles in the same time period for seven consecutive days in this traffic area. t i is the green light time during the process of the vehicle passing through the intersection in one direction of the intersection. T is the total sum of all green light and yellow light times for different directions of the same intersection. What is calculated is the probability of arriving at and passing through the intersection within time t i time. What is calculated is the probability of not being able to pass through the intersection within time t i time. What is calculated is the waiting time at the intersection.

10. An urban traffic route planning system, applied to the urban traffic route planning method according to any one of claims 1-9, characterized in that, The system includes a data integration unit, a path planning unit, and a path application unit; The data integration unit is configured to integrate the obtained large data set and perform processing, in-depth mining, and analysis to obtain traffic parameter data; The path planning unit is configured to use a multi-dimensional dynamic model to predict future traffic demands, and according to the prediction results, combine traffic parameter data, and apply a path planning algorithm to generate an optimal path planning scheme; The path application unit is configured to improve and adjust the path planning of the user during driving based on the optimal path planning scheme generated by the path planning unit; Among them, the path planning unit includes: A model construction module configured to construct a multi-dimensional dynamic model by combining time series analysis and a neural network model; A demand prediction module configured to input the large data set deeply mined and analyzed by the data integration unit into the multi-dimensional dynamic model. The multi-dimensional dynamic model uses the methods of weighted average and stacking to integrate the prediction results of the time series analysis and the neural network model, and outputs the prediction of the multi-dimensional dynamic model for future traffic demands; A solution generation module, configured to generate an optimal path planning solution according to a path planning algorithm and in combination with the prediction results of a demand prediction module; The path application unit includes: A path optimization module, configured to improve the path planning of a user during driving based on the generated path planning solution; A monitoring feedback module, configured to collect traffic data in real time through a deployed real-time traffic monitoring system, analyze the data, evaluate the implementation effect of the path planning solution through the analysis, and feedback the evaluation results to the staff at the terminal. The staff adjusts and optimizes the path planning solution in a timely manner according to the evaluation results.

Images