Intelligent path planning method and system based on dynamic road condition prediction

Abstract

The invention provides an intelligent path planning method and system based on dynamic road condition prediction. The method comprises the steps of firstly obtaining multi-source dynamic data of a target area; then, constructing a weather influence prediction model to predict weather influence parameters in a future time period; secondly, constructing a weather-traffic coupling model, and respectively establishing correlation models of corresponding precipitation, traffic flow density and average vehicle speed according to road types through historical data analysis, so as to estimate the traffic efficiency of each road section under a dynamic weather condition; and finally, generating a plurality of candidate paths according to the passing efficiency, screening out an alternative path set meeting a multi-target optimization condition from the candidate paths, performing simulation evaluation on the alternative path set, and determining an optimal path according to a simulation result. Compared with a traditional static path planning method, the method has the advantages that the responsiveness of a traffic system to meteorological disasters is remarkably improved, and predictable navigation service is provided for intelligent network connection vehicles.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, specifically to an intelligent route planning method and system based on dynamic road condition prediction. Background Technology

[0002] With the acceleration of urbanization and the popularization of intelligent transportation systems, the accuracy of path planning algorithms has become a key technical challenge for improving travel efficiency and safety. Traditional Dijkstra's and A-star algorithms, as classic optimal path search methods, rely on static weights (such as fixed distances or historical average travel times) as their core assumptions, but neglect the dynamic impact of sudden weather changes on traffic flow. For example, severe weather such as heavy rain, snowfall, and fog can lead to a decrease in road capacity (vehicle speed reduction of 30%-50%) and an increase in accident rates (the probability of traffic accidents increases by more than 40% in rainy and snowy weather), causing the actual travel time of a road segment to deviate significantly from the algorithm's preset weights.

[0003] While existing navigation systems can partially integrate real-time traffic flow data, they lack refined modeling of the coupling effect between weather factors and traffic flow, leading to planning results that deviate from the true optimal path and even exacerbating traffic congestion and safety hazards.

[0004] In recent years, with the deep integration of new technologies such as big data, vehicle networking, and artificial intelligence with the transportation industry, the research on optimal path generation has shown a trend of shifting from traditional traffic engineering methods to artificial intelligence methods represented by reinforcement learning. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing an intelligent route planning method and system based on dynamic traffic condition prediction.

[0006] The first aspect of this invention provides an intelligent route planning method based on dynamic traffic condition prediction, as detailed below:

[0007] Acquire multi-source dynamic data of the target area, wherein the multi-source dynamic data includes at least meteorological data and traffic flow data;

[0008] Furthermore, the meteorological data includes at least one of precipitation, temperature, humidity, and wind speed, and the traffic flow data includes vehicle density, average vehicle speed, and road capacity.

[0009] A weather impact prediction model is constructed based on the meteorological data to predict weather impact parameters in future periods;

[0010] Furthermore, the weather impact prediction model is a time-series prediction model based on deep learning, including a long short-term memory network or a bidirectional long short-term memory network.

[0011] Based on the traffic flow data and the weather impact parameters, a weather-traffic coupling model is constructed to estimate the traffic efficiency of each road segment under dynamic weather conditions.

[0012] Furthermore, the method of dynamically estimating the traffic efficiency of road segments by introducing weather impact parameters to modify the basic traffic flow map of the weather-traffic coupling model further includes:

[0013] The traffic flow basic diagram is a triangular basic diagram, whose parameters include free flow velocity, critical density and congestion density, and the weather impact parameters are used to correct the free flow velocity and / or critical density.

[0014] Based on the traffic efficiency, multiple candidate paths are generated, and a set of alternative paths that meet the multi-objective optimization conditions is selected from them.

[0015] Furthermore, the set of alternative paths is generated using a genetic algorithm with an elite retention strategy, wherein the genetic algorithm performs multi-objective optimization with the Pareto optimal solution as the objective.

[0016] Furthermore, the multi-objective optimization conditions include at least two of the following: minimizing travel time, minimizing congestion risk, or maximizing path reliability.

[0017] The set of alternative paths is evaluated through simulation, and the optimal path is determined based on the simulation results.

[0018] Furthermore, the simulation evaluation employs a traffic simulation tool, which dynamically loads weather impact parameters and real-time traffic flow data to update vehicle speeds.

[0019] Furthermore, the determination of the optimal path is based on at least one of the time cost, energy cost, or safety indicators output by the simulation.

[0020] A second aspect of the present invention provides an intelligent route planning system based on dynamic traffic condition prediction, comprising:

[0021] The multi-source data acquisition module is used to acquire meteorological and traffic flow data of the target area in real time.

[0022] The weather impact prediction module is communicatively connected to the multi-source data acquisition module and is used to predict weather impact parameters for future periods based on the meteorological data.

[0023] The weather-traffic coupling module is communicatively connected to the weather impact prediction module and the multi-source data acquisition module, and is used to dynamically estimate the traffic efficiency of each road segment based on the weather impact parameters and traffic flow data.

[0024] The route optimization module is communicatively connected to the weather-traffic coupling module and is used to generate a set of alternative routes that meet the multi-objective optimization conditions.

[0025] The simulation decision module is communicatively connected to the path optimization module and is used to perform simulation evaluation on the set of candidate paths and output the optimal path.

[0026] The beneficial effects of this invention are:

[0027] First, at the dynamic route planning level, this invention uses machine learning algorithms to finely divide the target area and construct a multi-dimensional prediction model, which can capture the future precipitation change trend and the dynamic evolution of traffic density in real time. Compared with traditional methods that rely only on historical data or static parameters, the foresight and accuracy of route selection are improved. This feature makes this invention particularly suitable for areas with frequent heavy rainfall, effectively avoiding waterlogged road sections and traffic congestion, and providing users with a safer and more efficient real-time route optimization solution.

[0028] Secondly, regarding algorithm efficiency, this invention innovatively combines static data with a genetic algorithm. By pre-generating k optimal candidate paths, it significantly reduces computational complexity and response time. Real-world testing shows that, under the same road network scale, the computational efficiency is improved compared to the traditional exhaustive search method, significantly reducing hardware resource requirements and making high-frequency path planning services highly feasible on existing computing devices. This technical solution balances prediction accuracy and computational efficiency, providing a novel solution for real-time path optimization in intelligent transportation systems. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the overall process of this application;

[0030] Figure 2 This is the basic traffic flow map for this application;

[0031] Figure 3 This application uses a partial road section in a certain district of Hangzhou as an example to illustrate the selection of points and edges.

[0032] Figure 4 This is a rendering of the road network map constructed for this application;

[0033] Figure 5 Flowchart of the genetic algorithm with elite retention strategy in this application;

[0034] Figure 6 This is the architecture of the LSTM precipitation prediction model in this application. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without inventive effort are within the scope of protection of this invention.

[0036] This application provides an intelligent route planning method and system based on dynamic traffic condition prediction. It relies on historical precipitation data to predict precipitation and uses precipitation, traffic density, and average vehicle speed as state variables in the regression model construction to achieve route planning that takes into account dynamic traffic conditions.

[0037] like Figure 1 As shown in the figure, this application provides an intelligent route planning method based on dynamic traffic condition prediction, including the following steps:

[0038] S1: Construction of a local precipitation prediction model;

[0039] Furthermore, step S1 is specifically implemented through the following sub-steps:

[0040] S11: Delineate the study area. Based on needs such as vehicle navigation and urban traffic management, a targeted study area is delineated, and the corresponding meteorological stations for that area are obtained.

[0041] S12: Meteorological Data Acquisition. Historical meteorological data for the target area will be collected from meteorological stations corresponding to the study area, including but not limited to multi-dimensional time-series data such as precipitation, temperature, humidity, air pressure, and wind speed.

[0042] S13: Meteorological Data Preprocessing. The collected meteorological data is preprocessed as follows: missing values ​​are filled using interpolation; duplicate entries are removed, retaining only the first record; outliers are identified using statistical methods (3σ principle or box plot), and their reasonableness is judged based on meteorological knowledge (e.g., a sudden increase in precipitation during a rainstorm may be a reasonable value), with outliers replaced by the mean of nearby time periods; Min-Max normalization is used to transform the data to the [0,1] interval to construct a high-quality input feature sequence.

[0043] S14: Deep Learning Model Construction. Construct a deep learning-based precipitation prediction model, such as LSTM. Figure 6 As shown, appropriate parameters for the number of neurons and time step are set to capture the temporal dependencies in meteorological data.

[0044] The input layer portion includes:

[0045] Set the time step parameter to 6-24 hours to capture local weather change characteristics; reduce the dimensionality to 5-8 key features through principal component analysis (PCA);

[0046] The hidden layer part includes:

[0047] 3-5 layers of stacked LSTM;

[0048] Input layer → Hidden layer 1: 128 neurons (to capture high-frequency fluctuations);

[0049] Hidden layer 1 → Hidden layer 2: 64 neurons (extracting mid-cycle patterns);

[0050] Hidden layer 2 → Hidden layer 3: 32 neurons (modeling long-term dependencies) to capture temporal dependencies in meteorological data.

[0051] Using bidirectional LSTM (Bi-LSTM) to fuse preceding and following time series information improves the ability to predict asymmetric precipitation events;

[0052] The output layer section includes:

[0053] An autoregressive structure is used to predict future precipitation data point by point.

[0054] S15: Training Model Parameters. The preprocessed multidimensional time-series feature sequences are input into the deep learning model for training. The backpropagation algorithm is used to adjust the model parameters and optimize the prediction loss function to obtain the optimal model weights.

[0055] S16: Deploy an online learning mechanism. Receive real-time weather station data and calculate prediction bias. When the average absolute error over five consecutive time slices exceeds a threshold, fine-tune the model parameters.

[0056] S17: Precipitation Prediction. A trained deep learning model is used to predict local precipitation over several future time steps, and the prediction results are output to support subsequent related applications.

[0057] S2: Based on the traffic flow basic map, a model is constructed to represent the relationship between precipitation, traffic density, and average traffic speed, in order to obtain a weather-traffic coupled model.

[0058] Furthermore, step S2 is specifically implemented through the following sub-steps:

[0059] S21: Traffic flow data collection. Within the study area, select all roads and collect their historical traffic flow data (vehicle density). Average vehicle speed ) and the corresponding weather conditions (precipitation) at that time ).

[0060] S22: Construction of basic traffic flow map for road segment. For example... Figure 2 As shown, using the data collected in S21, a basic triangular traffic flow diagram is constructed for each road segment. Free flow corresponds to the rising straight line on the left side of the diagram, indicating that under low traffic density, vehicles can travel freely without interference from other vehicles. Free flow speed. The slope of this section represents the average speed of vehicles in uncongested conditions, and the critical density. It is the end point of the free-flow section and the critical point where traffic flow transitions from a free state to a congested state. When the density exceeds... At this point, traffic flow enters a following state. The free-flow speed of each road segment is calculated. Congestion density Critical density , maximum traffic capacity Parameters such as these.

[0061] S23: Construction of a three-parameter relationship model based on the basic traffic flow map. In traditional traffic flow research, road segment speeds can be constructed based on data. Traffic density The correlation function. This method introduces precipitation data based on the basic traffic flow map. The amount of precipitation Traffic density As the independent variable, average vehicle speed As the dependent variable, multinomial regression is used to construct corresponding correlation functions according to road type:

[0062]

[0063] S24: Model parameter training. For each road within the region, the model parameters are trained using the corresponding historical traffic flow data from S21.

[0064] S25: Average speed calculation for road segments. Input the precipitation amount for each road segment. Traffic density This allows us to estimate the average speed of vehicles on the road segment. This further supports the calculation of travel time for road segments.

[0065] S3: Generate k Pareto optimal paths

[0066] Furthermore, step S3 is specifically implemented through the following sub-steps:

[0067] S31: Road Network Map Construction. Based on the study area defined in S11, the road network within the study area is saved in the form of a graph. Intersections are treated as points, and road segments as edges, such as... Figure 3 , Figure 4 As shown. Figure 4 To preserve Figure 3 A schematic diagram of the road network construction in the central part of the road network will be provided. Figure 4 Intersections marked with "ak" are connected by road segments, which are marked as road segments 1-15.

[0068] S32: Initial Definition of Road Network Weights. In the road network, the initial weight of each edge is defined as the average travel time under free-flow conditions. ,in Let be the length of the road segment, and u be the free-flow velocity of the road segment.

[0069] S33: Road network map weights are updated in real time. Based on the current traffic flow density and precipitation, and the average speed model of road segments constructed based on S25, the weights of each road segment are updated.

[0070] S34: Determine the starting and ending points of the path, and use a genetic algorithm with an elite retention strategy to generate k optimal alternative paths.

[0071] like Figure 5 As shown, firstly, based on the starting point and the ending point, a set of legal paths is randomly generated as the initial population. Each path consists of a series of nodes connecting the starting point to the ending point, ensuring the connectivity and legality of the path. After the selection, crossover, mutation, and update of the population, k paths located on the Pareto front are selected from the final population as the optimal path set under the multi-objective optimization balance.

[0072] S4: Perform detailed simulation and decision-making on k Pareto optimal paths.

[0073] Furthermore, step S4 is specifically implemented through the following sub-steps:

[0074] S41: Select a suitable traffic simulation tool to generate a basic road network. Use various intersections (such as crossroads, T-junctions, roundabouts, etc.), special functional nodes (such as expressway entrances and exits, locations where temporary traffic control measures are set up for construction activities), and road connection points as road nodes; use road segments as edges, and input the speed limit, traffic flow, and lane type of the road segments; input the traffic light timing.

[0075] S42: Construct route information. Add elements such as departure points and routes to specify the exact routes of the simulated vehicles.

[0076] S43: Establish a data server. Select a cloud server as needed.

[0077] S44: Implementation of dynamic vehicle speed setting. Real-time acquisition of predicted precipitation, traffic density, and road type data is used to predict speed. The prediction results are input into the data server built in S43, and the vehicle speed in the simulation is updated in real time based on the speed limit conditions.

[0078] S45: Determine the optimal path. Utilize the dynamic update method in S44 to perform a refined simulation of the k optimal paths, calculate the time cost of each of the k paths, and make a final decision based on its magnitude.

[0079] Based on the same concept as the above method embodiments, this application also proposes an intelligent route planning system based on dynamic traffic condition prediction, including:

[0080] The multi-source data acquisition module is used to acquire meteorological and traffic flow data of the target area in real time.

[0081] The weather impact prediction module is communicatively connected to the multi-source data acquisition module and is used to predict weather impact parameters for future periods based on the meteorological data.

[0082] The weather-traffic coupling module is communicatively connected to the weather impact prediction module and the multi-source data acquisition module, and is used to dynamically estimate the traffic efficiency of each road segment based on the weather impact parameters and traffic flow data.

[0083] The route optimization module is communicatively connected to the weather-traffic coupling module and is used to generate a set of alternative routes that meet the multi-objective optimization conditions.

[0084] The simulation decision module is communicatively connected to the path optimization module and is used to perform simulation evaluation on the set of candidate paths and output the optimal path.

[0085] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0086] The specific implementation paradigms described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An intelligent route planning method based on dynamic traffic condition prediction, characterized in that, Includes the following steps: Acquire multi-source dynamic data of the target area, wherein the multi-source dynamic data includes at least meteorological data and traffic flow data; A weather impact prediction model is constructed based on the meteorological data to predict weather impact parameters in future periods; Based on the traffic flow data and the weather impact parameters, a weather-traffic coupling model is constructed to estimate the traffic efficiency of each road segment under dynamic weather conditions. Based on the traffic efficiency, multiple candidate paths are generated, and a set of alternative paths that meet the multi-objective optimization conditions is selected from them. The set of alternative paths is evaluated through simulation, and the optimal path is determined based on the simulation results.

2. The intelligent route planning method based on dynamic traffic condition prediction according to claim 1, characterized in that, The meteorological data includes at least one of precipitation, temperature, humidity and wind speed, and the traffic flow data includes vehicle density, average vehicle speed and road capacity.

3. The intelligent route planning method based on dynamic traffic condition prediction according to claim 2, characterized in that, The weather impact prediction model is a time-series prediction model based on deep learning, including a long short-term memory network or a bidirectional long short-term memory network.

4. A smart route planning method based on dynamic traffic condition prediction according to any one of claims 1 to 3, characterized in that, The weather-traffic coupling model modifies the basic traffic flow map by introducing weather impact parameters to dynamically estimate the traffic efficiency of road segments.

5. The intelligent route planning method based on dynamic traffic condition prediction according to claim 4, characterized in that, The traffic flow basic diagram is a triangular basic diagram, and its parameters include free flow velocity, critical density and congestion density. The weather impact parameters are used to correct the free flow velocity and / or critical density.

6. The intelligent route planning method based on dynamic traffic condition prediction according to claim 1, characterized in that, The multi-objective optimization conditions include at least two of the following: minimizing travel time, minimizing congestion risk, or maximizing path reliability.

7. The intelligent route planning method based on dynamic traffic condition prediction according to claim 6, characterized in that, The set of alternative paths is generated by a genetic algorithm with an elite retention strategy, which performs multi-objective optimization with the Pareto optimal solution as the goal.

8. The intelligent route planning method based on dynamic traffic condition prediction according to claim 1, characterized in that, The simulation evaluation uses a traffic simulation tool, which dynamically loads weather impact parameters and real-time traffic flow data to update vehicle speeds.

9. A smart route planning method based on dynamic traffic condition prediction according to claim 1 or 8, characterized in that, The determination of the optimal path is based on at least one of the time cost, energy cost, or security indicators output by the simulation.

10. An intelligent route planning system based on dynamic traffic condition prediction, characterized in that, include: The multi-source data acquisition module is used to acquire meteorological and traffic flow data of the target area in real time. The weather impact prediction module is communicatively connected to the multi-source data acquisition module and is used to predict weather impact parameters for future periods based on the meteorological data. The weather-traffic coupling module is communicatively connected to the weather impact prediction module and the multi-source data acquisition module, and is used to dynamically estimate the traffic efficiency of each road segment based on the weather impact parameters and traffic flow data. The route optimization module is communicatively connected to the weather-traffic coupling module and is used to generate a set of alternative routes that meet the multi-objective optimization conditions. The simulation decision module is communicatively connected to the path optimization module and is used to perform simulation evaluation on the set of candidate paths and output the optimal path.

Images