Urban road traffic state prediction method based on macro-medium-micro hybrid simulation model, storage medium and equipment

Abstract

The invention discloses an urban road traffic state prediction method based on a macro-medium-micro hybrid simulation model, a storage medium and equipment, and the method comprises the steps: collecting historical electronic police data and traffic channelization data, inputting the historical electronic police data and traffic channelization data into a constructed macro-medium-micro integrated multi-level hybrid simulation model, and carrying out the simulation, obtaining a road traffic state in each time period and traffic flow data at each moment; constructing a traffic flow parameter prediction model based on an MVF-FA-LSTM algorithm, inputting the traffic flow data into the traffic flow parameter prediction model, and predicting the traffic flow and the passing speed of the road; and constructing a traffic state prediction model based on a fuzzy C-means clustering algorithm, inputting the predicted traffic flow and passing speed of the road into the traffic state prediction model, and predicting the traffic state of the road. According to the method, unification of microscopic, mesoscopic and macroscopic simulation models is realized, and road flow and passing speed prediction is realized, so that road traffic states are divided, and a traffic management department is guided to command and make decisions.

Description

Technical Field

[0001] The present invention belongs to the technical field of traffic simulation modeling, and in particular, relates to a method, storage medium and device for predicting urban road traffic status based on a macro-micro hybrid simulation model. Background Art

[0002] Traffic simulation technology is an effective means to reproduce the spatiotemporal evolution of traffic flow. It can help urban traffic management personnel simulate and analyze the evolution of traffic flow in a computer environment, deduce the impact of various planning and traffic management changes on traffic, and empower daily traffic management. However, based on the summary of many years of actual management, the current traditional traffic simulation still has the following shortcomings:

[0003] (1) Currently, most traffic simulations adopt a “node-arc” modeling structure, which requires “lane-to-lane connection” and cannot be automatically connected. This makes it difficult to construct three-dimensional roads such as overpasses and viaducts. In addition, modeling intersections with special phases requires complex signal configuration, which makes modeling difficult and inefficient.

[0004] (2) Traditional traffic simulation is mainly carried out separately at different levels according to different needs. Different traffic models are independent of each other and have no correlation. When the scope of traffic simulation is expanded, the vehicle path data in the traffic model needs to be manually estimated. There is a lack of fusion analysis combined with traffic volume data. The turning, dissipation, periodicity, and sudden behavior of vehicles between road sections cannot be accurately intervened, which affects the simulation results.

[0005] (3) Traditional traffic simulation mostly inputs historical status data into an existing road network and uses the result data output by the offline model for visualization. The simulation results have a significant lag and cannot build a full-time and space closed-loop simulation system from the past to the present and the future. As a result, the simulation is only used as an auxiliary analysis and decision-making tool, and cannot judge the future road network congestion trend. Summary of the invention

[0006] In response to the problems existing in the prior art, the present invention provides a method, storage medium and device for predicting urban road traffic status based on a macro-meso-micro hybrid simulation model, which realizes the unification of micro, meso and macro simulation models, and realizes the prediction of road flow and travel speed, thereby dividing road traffic status and guiding road traffic management.

[0007] To achieve the above technical objectives, the present invention adopts the following technical solution: a method for predicting urban road traffic status based on a macro-mid-micro hybrid simulation model, comprising the following steps:

[0008] Step S1, constructing a macro-, micro-integrated multi-level hybrid simulation model based on GIS data of urban traffic, traffic channelization data, electric police data, planning and construction data, and information control data;

[0009] Step S2: collect historical electric alarm data and traffic channelization data, input them into the constructed macro-medium-micro integrated multi-level hybrid simulation model, perform simulation, and obtain the road traffic status in each time period and the traffic flow data at each time;

[0010] Step S3: construct a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm, input the traffic flow data into the constructed traffic flow parameter prediction model, and predict the traffic flow and travel speed of the road;

[0011] Step S4: construct a traffic state prediction model based on the fuzzy C-means clustering algorithm, input the predicted traffic flow and travel speed of the road into the constructed traffic state prediction model, and predict the traffic state of the road.

[0012] Furthermore, step S1 includes the following sub-steps:

[0013] Step S1.1, using the "line connection" structure in the Aimsun simulation software to build a road network model with traffic channelization data, defining the GIS data in the road network model according to the GIS data, setting the road parameters of each section in the road network model, including: road type, road capacity and road speed limit, and adding electric police detection points in the road network model according to the electric police data;

[0014] Step S1.2, performing traffic signal timing at the intersection of the road network model according to the signal control data;

[0015] Step S1.3, dividing the traffic area according to the planning and construction data, and forming an OD travel chain between the traffic area and the roads in the road network model;

[0016] Step S1.4, input the travel volume of the traffic zone into the Aimsun simulation software to generate an OD travel matrix, perform multiplication or division on the OD travel matrix according to the proportion of urban traffic travel modes, generate OD travel matrices of different travel modes, realize the division of simulated traffic travel modes, and obtain a preliminary urban traffic simulation model;

[0017] Step S1.5: Use the electronic police detection points to collect travel flow, input it into the preliminary urban traffic simulation model, adjust the OD travel matrix of different travel modes, and distribute it to different roads to complete the construction of the macro-medium-micro integrated multi-level hybrid simulation model.

[0018] Furthermore, step S1.5 includes the following sub-steps:

[0019] Step S1.5.1, split the OD travel matrix of different travel modes in time sequence, adjust the traffic travel demand in the split OD travel matrix according to the travel flow collected by the electric police detection point, and consider the travel time between the OD travel chain and the electric police detection point to update the OD travel matrix;

[0020] Step S1.5.2: Divide the traffic travel time period into equal intervals, establish a dynamic user equilibrium model, and distribute the traffic travel to different roads.

[0021] Furthermore, step S3 includes the following sub-steps:

[0022] Step S3.1, adding the arithmetic mean filter MVF to the input gate and the forget gate of the LSTM respectively, and using the firefly algorithm FA to optimize the parameters of the LSTM with the arithmetic mean filter MVF added, to construct a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm;

[0023] Step S3.2: Input the traffic flow data into the constructed traffic flow parameter prediction model to predict the traffic flow and travel speed of the road.

[0024] Furthermore, step S3.1 includes the following sub-steps:

[0025] Step S3.1.1, collect historical traffic flow data in urban traffic, use the traffic flow and speed of the road at each moment in the historical traffic flow data as the training labels of LSTM, and calculate the arithmetic mean of all traffic flow data at each moment in the historical traffic flow data;

[0026] Step S3.1.2, the time window size, batch size and number of hidden layer units of LSTM are combined into firefly individual vectors, the firefly population size is set, and the firefly individual vectors are randomly initialized, and the light intensity absorption coefficient, maximum attraction factor and number of iterations are set;

[0027] Step S3.1.3, use the firefly individual vector to configure LSTM, when the historical traffic flow data is input into LSTM for processing, the input gate and forget gate of LSTM are filtered by the arithmetic mean of the traffic flow data at the corresponding moment, and the traffic flow and speed of the road at each moment are predicted;

[0028] Step S3.1.4, calculate the mean square error between the predicted traffic flow and speed of the road at each moment and the training label at the corresponding moment, and retain the firefly individual vector with the smallest mean square error;

[0029] Step S3.1.5: In the next iteration, update the positions of other firefly individual vectors, and repeat steps S3.1.3 to S3.1.4 for the firefly individual vectors with updated positions, compare the minimum mean square error in the current iteration with the mean square error retained in the previous iteration, and retain the firefly individual vector with the minimum mean square error;

[0030] Step S3.1.6: Repeat step S3.1.5 until the maximum number of iterations is reached to obtain the optimal firefly individual vector, which is used as the optimal parameter of LSTM to complete the construction of the traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm.

[0031] Furthermore, the process of updating the positions of other firefly vectors in step S3.1.5 is specifically as follows:

[0032] x i (t+1)=x i (t)+β ij [x j (t)-x i (t)]+αε

[0033] Among them, x i (t) represents the position of the individual vector of the i-th firefly at the t-th iteration, x j (t) represents the position of the firefly individual vector retained at the tth iteration, β ij represents the attraction between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, β0 represents the maximum attraction factor, γ represents the light intensity absorption coefficient, r ij represents the distance between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, d represents the dimension of the firefly individual vector, k represents the index of d, and x ik represents the position of the k-th dimension element in the i-th firefly individual vector at the t-th iteration, x jk represents the position of the k-th dimension element in the firefly individual retained at the t-th iteration, α represents the step size factor, and ε represents a random number generated by a Gaussian distribution.

[0034] Furthermore, step S4 includes the following sub-steps:

[0035] Step S4.1, taking the historical road traffic status as the initial cluster and initializing each cluster center;

[0036] Step S4.2, clustering the road traffic flow and traffic speed at each moment into cluster samples;

[0037] Step S4.3, calculating the Euclidean distance from each cluster sample to each cluster center, and assigning the cluster sample to the cluster with the smallest Euclidean distance;

[0038] Step S4.4, determining the membership of each cluster sample to each cluster according to the Euclidean distance, and combining all cluster samples in each cluster to update each cluster center;

[0039] Step S4.5, repeating steps S4.3-S4.4 until the maximum number of iterations is reached, obtaining the optimal cluster center and membership degree, and completing the construction of the traffic state prediction model;

[0040] Step S4.6: Input the predicted road traffic flow and travel speed at each moment into the traffic state prediction model to predict the traffic state of the corresponding road.

[0041] Furthermore, in step S4.4, the membership of each cluster sample to each cluster μ ab The calculation process is:

[0042]

[0043] Where C represents the number of cluster centers, c represents the index of C, and d ab represents the Euclidean distance from the ath cluster sample to the bth cluster center, d ac represents the Euclidean distance from the ath cluster sample to the cth cluster center, and m represents the fuzziness;

[0044] Each cluster center c c The update process is:

[0045]

[0046] Among them, x a represents the ath cluster sample, and n represents the total number of cluster samples belonging to cluster center c.

[0047] Furthermore, the constructed macro-, micro-integrated multi-level hybrid simulation model is visualized.

[0048] Furthermore, the present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the urban road traffic state prediction method based on the macro-micro hybrid simulation model.

[0049] Furthermore, the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the urban road traffic status prediction method based on the macro-medium-micro hybrid simulation model is implemented.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] (1) The urban road traffic state prediction method based on the macro-meso-micro hybrid simulation model of the present invention realizes the unification of micro, meso and macro simulation models through the macro-meso-micro integrated multi-level hybrid simulation model, makes up for the shortcomings of different simulation models, integrates different simulation models together, selects corresponding simulation models according to different needs of users, realizes accurate traffic simulation in different scenarios, breaks through the information islands between different models, strengthens the information association between micro, meso and macro simulation models, ensures the correlation and consistency between key model information, and realizes the association of different simulation models;

[0052] (2) The present invention realizes the prediction of road traffic flow and travel speed through the traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm, which can effectively solve the shortcomings of the traditional LSTM algorithm model, such as cumbersome parameter setting and long feedback time, thereby improving the accuracy and feasibility of traffic flow parameter prediction, and can effectively mine the time characteristic attributes of urban road traffic management, providing an accurate data basis for subsequent road traffic status prediction;

[0053] (3) The present invention predicts the traffic status of a road by constructing a traffic status prediction model based on the fuzzy C-means clustering algorithm. The traffic status prediction model constructed by the fuzzy C-means clustering algorithm has better generalization ability, improves the accuracy of road traffic status prediction, and is suitable for different urban road traffic scenarios. It then effectively provides the congestion evolution law of the road network in the future, helps the traffic management department to better formulate traffic management strategies, and improves the overall operation efficiency of the urban traffic system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a framework diagram of the urban road traffic state prediction method based on the macro-mid-micro hybrid simulation model of the present invention. DETAILED DESCRIPTION

[0055] The technical solution of the present invention is further explained below in conjunction with the accompanying drawings.

[0056] like Figure 1 The framework diagram of the urban road traffic state prediction method based on the macro-micro hybrid simulation model of the present invention includes the following steps:

[0057] Step S1: construct a macro-, micro-integrated multi-level hybrid simulation model based on GIS data of urban traffic, traffic channelization data, electric police data, planning and construction data, and signal control data; specifically, the following sub-steps are included:

[0058] Step S1.1, using the "line connection" structure in the Aimsun simulation software to connect multiple roads to form a way of connecting multiple roads, so as to accurately and quickly build a road network model, define the GIS data in the road network model according to the GIS data, set the road parameters of each section in the road network model, including: road type, road capacity and road speed limit, and add electric police detection points in the road network model according to the electric police data; among them, the road types are divided into expressways, trunk roads, secondary trunk roads and branch roads, and the road capacity and road speed limit are defined according to different road types;

[0059] Step S1.2, performing traffic signal timing at the intersection of the road network model according to the signal control data; specifically, according to the current release phases of different intersections in the established road network model, setting the phases in the road network model and binding different flow directions with the phases, so as to set the signal light cycles of different intersections and the green light durations of different phases according to the actual signal control data;

[0060] Step S1.3, dividing the traffic area according to the planning and construction data, and forming an OD travel chain between the traffic area and the roads in the road network model;

[0061] Step S1.4, encode and bind the cells in the road network model with the actual cells, generate the diagonal matrix of the travel volume of all cells with the help of Python, input it into the Aimsun simulation software, generate the OD travel matrix, multiply or divide the OD travel matrix according to the proportion of urban traffic travel modes, generate the OD travel matrix of different travel modes, such as motor vehicles, non-motor vehicles, public transportation and other travel modes, realize the division of simulated traffic travel modes, and obtain a preliminary urban traffic simulation model;

[0062] Step S1.5: Use Python to associate the electric police points in the road network model with the actual electric police points provided by the traffic management department, use the electric police detection points to collect travel flow, input it into the preliminary urban traffic simulation model, adjust the OD travel matrix of different travel modes, and distribute it to different roads, completing the construction of the macro-mid-micro integrated multi-level hybrid simulation model, including the following sub-steps:

[0063] Step S1.5.1, split the OD travel matrix of different travel modes in time sequence, adjust the traffic travel demand in the split OD travel matrix according to the travel flow collected by the electric police detection point, and consider the travel time between the OD travel chain and the electric police detection point, update the OD travel matrix, so that the updated OD travel matrix is ​​closer to the real travel data;

[0064] Step S1.5.2: Divide the traffic travel time periods into equal intervals, establish a dynamic user equilibrium model, and distribute traffic travel to different roads, thereby reducing travel costs.

[0065] After the above-mentioned macro-meso-micro integrated multi-level hybrid simulation model is built, you can select macro, meso and micro areas according to actual business needs, and simulate the selected areas. The macro-meso-micro integrated multi-level hybrid simulation model realizes the unification of micro, meso and macro simulation models, avoids multiple reconstructions of traffic simulation models for different needs, makes up for the shortcomings of different simulation models, integrates different simulation models together, selects corresponding simulation models according to different user needs, realizes accurate traffic simulation in different scenarios, breaks through the information islands between different models, strengthens the information association between micro, meso and macro simulation models, ensures the correlation and consistency between key model information, and realizes the association of different simulation models.

[0066] Step S2, collect historical electric police data and traffic channelization data, input the constructed macro-, micro-integrated multi-level hybrid simulation model, perform simulation, and obtain the mode of road traffic in each time period and the traffic flow data at each time. The modes of road traffic include 12 different traffic modes, including weekday flat peak, weekday morning peak, weekday evening peak, weekday night, weekend flat peak, weekend morning peak, weekend evening peak, weekend night, holiday flat peak, holiday morning peak, holiday evening peak, and holiday night. For these 12 traffic modes, their traffic status can be judged. There are 6 levels of traffic status, among which level A means that the vehicle operation is completely unimpeded. , the intersection delay is minimal; Class B indicates a reasonable unobstructed condition, that is, the vehicle's ability to change lanes during driving is less restricted, and the intersection delay is small; Class C indicates a stable operating state. Compared with Class B, the vehicle's ability to change lanes during driving is more restricted, but it can still maintain stable operation; Class D indicates a critical operating state, that is, a slight increase in traffic flow will cause traffic congestion, and the vehicle's ability to change lanes during driving is significantly restricted; Class E indicates an unstable state, the maneuverability in the traffic flow is severely restricted, the vehicle's ability to change lanes during driving is seriously affected, and traffic congestion occurs frequently; F indicates a blocked state, that is, the traffic flow is completely stagnant and vehicles cannot move.

[0067] Step S3, construct a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm, input the traffic flow data into the constructed traffic flow parameter prediction model, and predict the traffic flow and travel speed of the road. The traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm reduces the interference of uncertain factors by introducing the arithmetic mean filter MVF. At the same time, the LSTM parameter optimization is performed through the firefly algorithm FA, which can effectively solve the shortcomings of the traditional LSTM parameter setting being cumbersome and the feedback time being long, thereby improving the accuracy and feasibility of traffic flow parameter prediction, and can effectively mine the time characteristic attributes of urban road traffic management, and provide an accurate data basis for subsequent road traffic status prediction; specifically, it includes the following sub-steps:

[0068] Step S3.1, in practical applications, the effect of LSTM prediction is often affected by uncertain factors such as weather, traffic control, and traffic accidents, and is prone to large prediction errors. In order to reduce the interference of uncertain factors on the traffic flow parameter prediction model, the arithmetic mean filter MVF is added to the input gate and the forget gate in the LSTM, so as to filter the noise in the LSTM prediction process, and the firefly algorithm FA is used to optimize the parameters of the LSTM with the arithmetic mean filter MVF added, and a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm is constructed; specifically,

[0069] Step S3.1.1, collect historical traffic flow data in urban traffic, use the traffic flow and speed of the road at each moment in the historical traffic flow data as the training labels of LSTM, and calculate the arithmetic mean of all traffic flow data at each moment in the historical traffic flow data;

[0070] Step S3.1.2, the time window size, batch size and number of hidden layer units of LSTM are combined into firefly individual vectors, the firefly population size is set, and the firefly individual vectors are randomly initialized, and the light intensity absorption coefficient, maximum attraction factor and number of iterations are set;

[0071] Step S3.1.3, use the firefly individual vector to configure LSTM, when the historical traffic flow data is input into LSTM for processing, the input gate and forget gate of LSTM are filtered by the arithmetic mean of the traffic flow data at the corresponding moment, and the traffic flow and speed of the road at each moment are predicted;

[0072] Step S3.1.4, calculate the mean square error between the predicted traffic flow and speed of the road at each moment and the training label at the corresponding moment, and retain the firefly individual vector with the smallest mean square error;

[0073] Step S3.1.5: In the next iteration, update the positions of other firefly individual vectors, and repeat steps S3.1.3 to S3.1.4 for the firefly individual vectors with updated positions, compare the minimum mean square error in the current iteration with the mean square error retained in the previous iteration, and retain the firefly individual vector with the minimum mean square error;

[0074] The process of updating the positions of other firefly vectors in the present invention is specifically as follows:

[0075] x i (t+1)=x i (t)+β ij [x j (t)-x i (t)]+αε

[0076] Among them, x i (t) represents the position of the individual vector of the i-th firefly at the t-th iteration, x j (t) represents the position of the firefly individual vector retained at the tth iteration, β ij represents the attraction between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, β0 represents the maximum attraction factor, γ represents the light intensity absorption coefficient, r ij represents the distance between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, d represents the dimension of the firefly individual vector, k represents the index of d, and x ik represents the position of the k-th dimension element in the i-th firefly individual vector at the t-th iteration, x jk represents the position of the k-th dimension element in the firefly individual retained at the t-th iteration, α represents the step size factor, and ε represents a random number generated by a Gaussian distribution.

[0077] Step S3.1.6: Repeat step S3.1.5 until the maximum number of iterations is reached to obtain the optimal firefly individual vector, which is used as the optimal parameter of LSTM to complete the construction of the traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm.

[0078] Step S3.2: Input the traffic flow data into the constructed traffic flow parameter prediction model to predict the traffic flow and travel speed of the road.

[0079] In order to illustrate the reliability of the traffic flow parameter prediction model constructed by the present invention, 24 days of historical data of a certain intersection were selected as a training set, and 6 days of data were selected as a test set. The traffic flow parameter prediction model constructed by the present invention was optimized by the Adam function, and the number of iterations was set to 500 times, the size of the firefly population was 50, the maximum attraction factor was 1, and the light intensity absorption coefficient was 1. The traffic flow prediction model constructed by the present invention was compared with the traditional LSTM for traffic flow prediction. It can be seen that the traffic flow predicted by the traffic flow parameter prediction model constructed by the present invention has an average absolute error MAE of 8.94% and an average absolute percentage error MAPE of 1.54%, while the traffic flow predicted by the traditional LSTM has an average absolute error MAE of 19.93% and an average absolute percentage error MAPE of 6.93%, indicating that the traffic flow parameter prediction model constructed by the present invention has higher accuracy and reliability when predicting traffic flow.

[0080] Step S4: construct a traffic state prediction model based on the fuzzy C-means clustering algorithm, input the predicted road traffic flow and travel speed into the constructed traffic state prediction model, and predict the traffic state of the road. The traffic state prediction model constructed by the fuzzy C-means clustering algorithm has better generalization ability, improves the accuracy of road traffic state prediction, and is suitable for different urban road traffic scenes. Then, it effectively provides the congestion evolution law of the road network at future moments, helps the traffic management department to better formulate traffic management strategies, and improves the overall operation efficiency of the urban traffic system; specifically, it includes the following sub-steps:

[0081] Step S4.1, taking the historical road traffic status as the initial cluster and initializing each cluster center;

[0082] Step S4.2, clustering the road traffic flow and traffic speed at each moment into cluster samples;

[0083] Step S4.3, calculating the Euclidean distance from each cluster sample to each cluster center, and assigning the cluster sample to the cluster with the smallest Euclidean distance;

[0084] Step S4.4, determining the membership of each cluster sample to each cluster according to the Euclidean distance, and combining all cluster samples in each cluster to update each cluster center;

[0085] The membership degree μ of each cluster sample to each cluster in the present invention is ab The calculation process is:

[0086]

[0087] Where C represents the number of cluster centers, c represents the index of C, and d ab represents the Euclidean distance from the ath cluster sample to the bth cluster center, dac represents the Euclidean distance from the ath cluster sample to the cth cluster center, and m represents the fuzziness;

[0088] Each cluster center c c The update process is:

[0089]

[0090] Among them, x a represents the ath cluster sample, and n represents the total number of cluster samples belonging to cluster center c.

[0091] Step S4.5, repeating steps S4.3-S4.4 until the maximum number of iterations is reached, obtaining the optimal cluster center and membership degree, and completing the construction of the traffic state prediction model;

[0092] Step S4.6: Input the predicted road traffic flow and travel speed at each moment into the traffic state prediction model to predict the traffic state of the corresponding road.

[0093] The urban road traffic state prediction method based on the macro-micro hybrid simulation model of the present invention can be simulated according to the real traffic flow data through the macro-micro integrated multi-level hybrid simulation model, and the prediction of road traffic flow and travel speed can be realized through the constructed traffic flow parameter prediction model, and the traffic state prediction model can be used to predict the road traffic state. At the same time, the predicted road traffic flow and travel speed can be input into the macro-micro integrated multi-level hybrid simulation model for simulation, so as to verify whether the predicted road traffic state is correct. In addition, for the predicted traffic state at the DF level, after the traffic management department adjusts the traffic management strategy in time, the effect of the adjusted traffic management strategy can be evaluated through the simulation of the macro-micro integrated multi-level hybrid simulation model, thereby realizing the closed loop of traffic management work.

[0094] In a technical solution of the present invention, the constructed macro-mid-micro integrated multi-level hybrid simulation model is visualized.

[0095] In a technical solution of the present invention, a computer-readable storage medium is also provided, storing a computer program, wherein the computer program enables a computer to execute the urban road traffic state prediction method based on the macro-micro hybrid simulation model.

[0096] In one technical solution of the present invention, an electronic device is also provided, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the urban road traffic status prediction method based on the macro-medium-micro hybrid simulation model is implemented.

[0097] In the embodiments disclosed in the present application, the computer storage medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. The computer storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. More specific examples of computer storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0098] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0099] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for predicting urban road traffic status based on a macro-middle-micro hybrid simulation model, characterized in that: The steps include: Step S1, constructing a macro-, micro-integrated multi-level hybrid simulation model based on GIS data of urban traffic, traffic channelization data, electric police data, planning and construction data, and information control data; Step S2: collect historical electric alarm data and traffic channelization data, input them into the constructed macro-medium-micro integrated multi-level hybrid simulation model, perform simulation, and obtain the road traffic status in each time period and the traffic flow data at each time; Step S3: construct a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm, input the traffic flow data into the constructed traffic flow parameter prediction model, and predict the traffic flow and travel speed of the road; Step S4: construct a traffic state prediction model based on the fuzzy C-means clustering algorithm, input the predicted traffic flow and travel speed of the road into the constructed traffic state prediction model, and predict the traffic state of the road.

2. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 1 is characterized in that: Step S1 includes the following sub-steps: Step S1.1, using the "line connection" structure in the Aimsun simulation software to build a road network model with traffic channelization data, defining the GIS data in the road network model according to the GIS data, setting the road parameters of each section in the road network model, including: road type, road capacity and road speed limit, and adding electric police detection points in the road network model according to the electric police data; Step S1.2, performing traffic signal timing at the intersection of the road network model according to the signal control data; Step S1.3, dividing the traffic area according to the planning and construction data, and forming an OD travel chain between the traffic area and the roads in the road network model; Step S1.4, input the travel volume of the traffic zone into the Aimsun simulation software to generate an OD travel matrix, perform multiplication or division on the OD travel matrix according to the proportion of urban traffic travel modes, generate OD travel matrices of different travel modes, realize the division of simulated traffic travel modes, and obtain a preliminary urban traffic simulation model; Step S1.5: Use the electronic police detection points to collect travel flow, input it into the preliminary urban traffic simulation model, adjust the OD travel matrix of different travel modes, and distribute it to different roads to complete the construction of the macro-medium-micro integrated multi-level hybrid simulation model.

3. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 2 is characterized in that: Step S1.5 includes the following sub-steps: Step S1.5.1, split the OD travel matrix of different travel modes in time sequence, adjust the traffic travel demand in the split OD travel matrix according to the travel flow collected by the electric police detection point, and consider the travel time between the OD travel chain and the electric police detection point to update the OD travel matrix; Step S1.5.2: Divide the traffic travel time period into equal intervals, establish a dynamic user equilibrium model, and distribute the traffic travel to different roads.

4. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 1 is characterized in that: Step S3 includes the following sub-steps: Step S3.1, adding the arithmetic mean filter MVF to the input gate and the forget gate of the LSTM respectively, and using the firefly algorithm FA to optimize the parameters of the LSTM with the arithmetic mean filter MVF added, to construct a traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm; Step S3.2: Input the traffic flow data into the constructed traffic flow parameter prediction model to predict the traffic flow and travel speed of the road.

5. The method for predicting urban road traffic status based on a macro-middle-micro hybrid simulation model according to claim 4 is characterized in that: Step S3.1 includes the following sub-steps: Step S3.1.1, collect historical traffic flow data in urban traffic, use the traffic flow and speed of the road at each moment in the historical traffic flow data as the training labels of LSTM, and calculate the arithmetic mean of all traffic flow data at each moment in the historical traffic flow data; Step S3.1.2, the time window size, batch size and number of hidden layer units of LSTM are combined into firefly individual vectors, the firefly population size is set, and the firefly individual vectors are randomly initialized, and the light intensity absorption coefficient, maximum attraction factor and number of iterations are set; Step S3.1.3, use the firefly individual vector to configure LSTM, when the historical traffic flow data is input into LSTM for processing, the input gate and forget gate of LSTM are filtered by the arithmetic mean of the traffic flow data at the corresponding moment, and the traffic flow and speed of the road at each moment are predicted; Step S3.1.4, calculate the mean square error between the predicted traffic flow and speed of the road at each moment and the training label at the corresponding moment, and retain the firefly individual vector with the smallest mean square error; Step S3.1.5: In the next iteration, update the positions of other firefly individual vectors, and repeat steps S3.1.3 to S3.1.4 for the firefly individual vectors with updated positions, compare the minimum mean square error in the current iteration with the mean square error retained in the previous iteration, and retain the firefly individual vector with the minimum mean square error; Step S3.1.6: Repeat step S3.1.5 until the maximum number of iterations is reached to obtain the optimal firefly individual vector, which is used as the optimal parameter of LSTM to complete the construction of the traffic flow parameter prediction model based on the MVF-FA-LSTM algorithm.

6. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 5 is characterized in that: The process of updating the positions of other firefly vectors in step S3.1.5 is as follows: x i (t+1)=x i (t)+β ij [x j (t)-x i (t)]+αε Among them, x i (t) represents the position of the individual vector of the i-th firefly at the t-th iteration, x j (t) represents the position of the firefly individual vector retained at the tth iteration, β ij represents the attraction between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, β0 represents the maximum attraction factor, γ represents the light intensity absorption coefficient, r ij represents the distance between the i-th firefly individual vector and the retained firefly individual vector at the t-th iteration, d represents the dimension of the firefly individual vector, k represents the index of d, and x ik represents the position of the k-th dimension element in the i-th firefly individual vector at the t-th iteration, x jk represents the position of the k-th dimension element in the firefly individual retained at the t-th iteration, α represents the step size factor, and ε represents a random number generated by a Gaussian distribution.

7. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 1 is characterized in that: Step S4 includes the following sub-steps: Step S4.1, taking the historical road traffic status as the initial cluster and initializing each cluster center; Step S4.2, clustering the road traffic flow and traffic speed at each moment into cluster samples; Step S4.3, calculating the Euclidean distance from each cluster sample to each cluster center, and assigning the cluster sample to the cluster with the smallest Euclidean distance; Step S4.4, determining the membership of each cluster sample to each cluster according to the Euclidean distance, and combining all cluster samples in each cluster to update each cluster center; Step S4.5, repeating steps S4.3-S4.4 until the maximum number of iterations is reached, obtaining the optimal cluster center and membership degree, and completing the construction of the traffic state prediction model; Step S4.6: Input the predicted road traffic flow and travel speed at each moment into the traffic state prediction model to predict the traffic state of the corresponding road.

8. The method for predicting urban road traffic status based on a macro-micro hybrid simulation model according to claim 6 is characterized in that: The membership degree μ of each cluster sample to each cluster in step S4.4 ab The calculation process is: Where C represents the number of cluster centers, c represents the index of C, and d ab represents the Euclidean distance from the ath cluster sample to the bth cluster center, d ac represents the Euclidean distance from the ath cluster sample to the cth cluster center, and m represents the fuzziness; Each cluster center c c The update process is: Among them, x a represents the ath cluster sample, and n represents the total number of cluster samples belonging to cluster center c.

9. The method for predicting urban road traffic status based on a macro-middle-micro hybrid simulation model according to claim 1 is characterized in that: The constructed macro-, micro-integrated multi-level hybrid simulation model is visualized.

10. A computer-readable storage medium storing a computer program, characterized in that: The computer program enables the computer to execute the urban road traffic state prediction method based on the macro-micro hybrid simulation model as described in any one of claims 1 to 9.

11. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for predicting urban road traffic status based on a macro-micro hybrid simulation model as described in any one of claims 1 to 9 is implemented.

Images