Travel traffic road condition assessment planning system based on fusion processing

Abstract

The invention discloses a travel traffic road condition evaluation planning system based on fusion processing. The travel traffic road condition evaluation planning system comprises a multi-source data acquisition engine, a data fusion preprocessing unit, a road condition state fusion evaluation unit, a short-time prediction evolution unit, an intelligent path planning unit and a coordination signal optimization unit, according to the method, historical and real-time dual-mode prediction verification channel and regional coordination signal control feedback is established, a road network state evolution trend decoupling stability quantitative model is used for driving a planning decision, road average delay in peak periods in typical urban road network tests is decreased, the traffic capacity of main roads is improved, the response time of extreme meteorological events is shortened, and the reliability of the urban road network is improved. The road condition evaluation precision is improved through heterogeneous data fusion and dynamic rule engine coupling, a multi-source data acquisition engine is coupled with environmental data in real time, and signal noise interference is inhibited through a credibility weighting mechanism in a data fusion preprocessing unit.

Description

Technical Field

[0001] This invention belongs to the field of traffic control systems, and more specifically, relates to a travel traffic condition assessment and planning system based on fusion processing. Background Technology

[0002] In the field of traditional traffic control, mainstream systems adopt either a fixed-plan mode or a perception-based adjustment mode based on real-time feedback of local data. The former pre-sets daytime traffic light phase timing schemes, while the latter collects traffic flow parameters such as vehicle volume and speed through intersection cameras or loop detectors, triggering green light timings at individual nodes. Some systems have evolved to regional control architectures, attempting to optimize phase differences between adjacent intersection groups. The basic data of these systems typically relies on vehicle operation data from a single external source, as well as real-time data connected to license plate recognition devices.

[0003] The existing system is not capable of fusion and processing multi-source heterogeneous data, which is mainly reflected in three aspects: the relationship between actual influencing factors such as meteorological and environmental changes, temporary construction events, and sudden accidents and key links such as traffic modality spatial transmission efficiency is lacking system input, which makes the overall road condition assessment deviate from the current situation.

[0004] The traffic condition evolution prediction mechanism relies too heavily on historical pattern matching and cannot compensate for the disruption of the model when traffic flow changes abruptly. The disconnect between the path planning unit and the signal control unit submodule leads to the inclusion of real-time guidance loopholes in the recommended induced path while signal guidance fails, reducing the overall system operating efficiency and congestion avoidance effectiveness, and consequently affecting the overall experience of the travel control system and user trust in the credibility system. Therefore, we propose a travel traffic condition assessment and planning system based on fusion processing to address these shortcomings. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a traffic condition assessment and planning system based on fusion processing. This system establishes a historical and real-time dual-mode prediction verification channel and regional coordinated signal control feedback. It uses a decoupled stability quantification model of road network state evolution trend to drive planning decisions. In typical urban road network tests, the average road delay during peak hours is reduced, the capacity of main roads is improved, and the response time to extreme weather events is shortened.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A travel traffic condition assessment and planning system based on fusion processing includes:

[0008] Multi-source data acquisition engine, data fusion preprocessing unit, road condition fusion assessment unit, short-term prediction evolution unit, intelligent route planning unit, and coordinated signal optimization unit;

[0009] The traffic condition fusion assessment unit, based on the fused data output by the data fusion preprocessing unit, applies rule models and dynamic thresholds to determine the smooth, slow, and congested levels of regional road segments in real time, and estimates the congestion index and transit time, generating a timestamped traffic condition map. The short-term prediction evolution unit connects the traffic condition map unit and the historical database unit, combining real-time situation and historical patterns, and uses time series analysis to predict the traffic flow evolution trend of key road segments and road networks within a specified future time window. The intelligent route planning unit, based on the user's input of the travel origin, destination, preferences, real-time traffic condition map, and prediction evolution results, integrates and calculates to generate one or more congestion avoidance optimization route planning suggestions.

[0010] The output of the multi-source data acquisition engine is directly input to the data fusion preprocessing unit. The standard data output of the data fusion preprocessing unit drives the traffic condition fusion evaluation unit. The real-time status information of the traffic condition fusion evaluation unit is input to the short-term prediction evolution unit and the intelligent path planning unit, respectively. The prediction results of the short-term prediction evolution unit are also input to the intelligent path planning unit. The congestion situation information of the traffic condition fusion evaluation unit and the short-term prediction evolution unit are simultaneously input to the coordination signal optimization unit. The control signal of the coordination signal optimization unit acts on the urban traffic signal network. The induced path generated by the intelligent path planning unit can be fed back to the overall road network status evaluation system.

[0011] Preferably, the multi-level discrimination mechanism of the road condition fusion evaluation unit establishes a dynamic rule base containing a basic road segment traffic capacity matrix, and inputs standardized data into the hierarchical rule engine:

[0012] The first level determines the basic congestion level based on the ratio of real-time occupancy rate to the baseline saturation flow rate.

[0013] The second layer introduces the travel time index calculated from floating car trajectory data to perform topological correlation analysis on the synchronization status of adjacent road segments;

[0014] The third layer integrates meteorological visibility data with construction event coordinates to calculate the traffic capacity reduction for areas affected by the environment; the final output includes a dynamic traffic map containing road congestion index, confidence markers, and the distribution of affected lanes.

[0015] Preferably, the multi-mode prediction architecture of the short-term prediction evolution unit includes a historical pattern matching subunit and a real-time situational inference subunit;

[0016] The historical pattern matching subunit retrieves traffic flow curves for similar scenarios from the historical database based on date type and weather characteristics; the real-time situation inference unit receives real-time congestion boundary information from the road condition map and analyzes the traffic propagation path by combining the floating car trajectory vector direction.

[0017] A cross-validation mechanism is activated when conflicts are predicted: when the deviation between the real-time detected queue length growth rate and the historical pattern exceeds a threshold, the micro-inference results based on the vehicle car-following propagation model are used to generate a future road network state evolution map with probability confidence intervals.

[0018] Preferably, the congestion avoidance decision model of the intelligent path planning unit is pre-set with a road network topology connectivity matrix and a multi-objective cost function library, and after receiving user-defined cost weight parameters, it converts the real-time traffic map into a dynamic travel time weighted network.

[0019] Based on the improved heuristic search algorithm, the future multi-time road network status provided by the short-term prediction evolution unit is spatiotemporally unfolded, and the estimated travel time of each path at different times after departure is calculated.

[0020] Apply a time penalty coefficient to paths passing through predicted congestion zones to generate a set of congestion avoidance route schemes that include the range of estimated travel time fluctuations and energy consumption comparison indicators.

[0021] Preferably, the adaptive learning architecture of the short-term prediction evolution single is as follows:

[0022] Deploy a closed-loop verification module for prediction results to send the time series error values ​​between historical prediction data and actual detection data back to the pattern matching library;

[0023] Establish feature labels for special scenarios where prediction deviation exceeds a threshold, including scenarios such as sudden increases in rainfall and traffic flow after large events;

[0024] When real-time environmental parameters are detected to match specific scenario labels, the system automatically switches to an emergency prediction model trained using reinforcement learning.

[0025] Preferably, the intelligent route planning unit is configured with a group user route coordination interface to receive batch route requests from the logistics dispatch center;

[0026] Establish a shared road network load status matrix and mark the occupied path capacity in real time during the planning process; when multiple planned paths are detected to overlap in key bottleneck sections, staggered scheduling is carried out based on the prediction through time windows to generate a set of collaborative path schemes containing suggested departure time offsets.

[0027] Preferably, the road condition fusion assessment unit uses a dynamic congestion index calculation formula for discrimination:

[0028] ,in, Let be the fused congestion index of road segment i at time t. Real-time occupancy rate of road segments This refers to the saturation occupancy rate of road sections. The average speed of the road segment. This is the minimum speed limit for the section of road. Weather influencing factors The density of construction events is represented by α, β, γ, and δ, which are adaptive weighting coefficients. By combining exponential and logarithmic functions nonlinearly, the coupling effect between high occupancy and low vehicle speed is enhanced, transforming sudden weather and construction factors into superimposed quantitative impacts, and achieving effective aggregation of multi-source heterogeneous results.

[0029] Preferably, the predictive evolution calculation formula for traffic flow mutation identification in short-term predictive evolution units is as follows:

[0030] ,in, This represents the probability of sudden changes in traffic flow. Let τ be the length of the vehicle queue during a similar historical scenario. This is the queue length for real-time detection. To predict the rate of change of acceleration within a time window, λ is a road grade correction factor, with 0.8 for highways and 1.2 for urban roads. The deviation between historical patterns and real-time conditions is quantified by an exponential decay function, and the inflection point of traffic flow changes is predicted by combining the rate of change of acceleration, thereby improving the sensitivity to capturing high-volume impact events.

[0031] Preferably, the multi-source data acquisition engine acquires heterogeneous traffic flow data and environmental data from traffic cameras, floating car GPS, electronic sensors, roadside units, and historical databases in real time;

[0032] The multi-source data acquisition engine accesses real-time precipitation intensity data provided by traffic and meteorological radar stations, construction control information released by road maintenance departments, and millimeter-wave radar traffic monitoring devices deployed at urban expressway ramps. The engine has a built-in data quality verification submodule that marks abnormal sampling frequencies of floating car GPS data or broken frames in camera video streams and activates backup sensor nodes to supplement data acquisition. Environmental data acquisition is extended to road surface temperature and visibility indicators, and it is linked with the traffic incident alarm platform to capture information on traffic accidents and sudden control events in real time.

[0033] The technical effects and advantages of this invention are as follows: Compared with the prior art, the traffic condition assessment and planning system based on fusion processing provided by this invention improves the accuracy of traffic condition assessment by coupling heterogeneous data fusion with a dynamic rule engine. The multi-source data acquisition engine is coupled with environmental data in real time. In the data fusion preprocessing unit, a credibility weighting mechanism is used to suppress signal noise interference. The traffic condition fusion assessment unit performs capacity reduction calculation on environmental data based on the road segment capacity matrix and dynamically corrects the weight value of the influence of weather visibility on traffic flow.

[0034] In addition, the historical pattern matching subunit for short-term prediction evolution automatically retrieves historical flow curves under similar meteorological conditions, while the real-time situational simulation subunit analyzes the convoy vector propagation direction based on road condition maps. The dual-data input cross-validation module prioritizes micro-simulation based on the vehicle following principle when the flow propagation rate analyzed by the floating car trajectory deviates from the historical pattern. This mechanism effectively corrects prediction drift problems caused by sudden events during peak hours. Attached Figure Description

[0035] Figure 1 This is a diagram of the travel traffic condition assessment and planning system based on fusion processing, as described in this invention. Detailed Implementation

[0036] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. 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.

[0037] This invention provides, for example Figure 1 The traffic condition assessment and planning system based on fusion processing shown is used in traffic control systems to achieve accurate traffic condition assessment and proactive optimization based on a multi-technology fusion and collaborative mechanism. It constructs a closed-loop control chain of heterogeneous data fusion, accurate state assessment, evolution trend verification, optimization command execution, planning scheme output, and road network state feedback correction. In typical urban road network tests, the average road delay during peak hours is reduced, thereby improving the capacity of main roads and reducing the response time to extreme weather events.

[0038] The system specifically includes:

[0039] The multi-source data acquisition engine acquires heterogeneous traffic flow data in real time from traffic cameras, floating car GPS, electronic sensors, roadside units, and historical databases, including flow speed, occupancy, and event information; and environmental data, including weather and lighting.

[0040] It should be noted that the optimized structure of the multi-source data acquisition engine further integrates real-time precipitation intensity data provided by traffic and meteorological radar stations, construction control information released by road maintenance departments, and millimeter-wave radar traffic monitoring devices deployed at urban expressway ramps.

[0041] The engine has a built-in data quality verification submodule that marks abnormal sampling frequency of floating car GPS data or broken frames in camera video streams and activates backup sensor nodes to supplement data collection; environmental data collection is extended to road surface temperature and visibility indicators, and it is linked with the traffic incident alarm platform through a dedicated communication channel to capture information on traffic accidents or sudden traffic control events in real time.

[0042] The data fusion preprocessing unit receives data from the engine and performs spatiotemporal alignment, noise filtering, and outlier cleaning, fusing heterogeneous data into a unified, standardized spatiotemporal sequence dataset. Further, the operational architecture of the data fusion preprocessing unit is as follows:

[0043] It includes interconnected spatiotemporal alignment subunits, dynamic noise filtering subunits, and confidence weighting subunits; the spatiotemporal alignment subunit maps the floating car's GPS coordinates to the road link topology grid where the electronic sensors are located, and uses time window sliding interpolation for non-equal interval sampling data;

[0044] The dynamic noise filtering subunit generates an adaptive filtering threshold based on the road segment historical data feature library to eliminate abnormal vehicle speed values ​​caused by satellite positioning drift; the confidence weighting subunit assigns differentiated fusion weights to camera traffic data and coil occupancy data according to sensor type and historical accuracy to generate a standardized data stream with confidence labels.

[0045] Secondly, the data fusion preprocessing unit has a special processing procedure for construction event data: creating a spatiotemporal impact model of construction control events, mapping the coordinates of the construction area, lane closure scheme, and planned duration to a road network capacity attenuation matrix; for temporary construction sign information identified through camera video streams, initiating a manual verification channel to send a confirmation request to the road administration platform; after verification, updating the road topology in real time, embedding lane reduction information into the road attribute fields of the standardized data stream.

[0046] The road condition fusion assessment unit, based on the fused data output by the data fusion preprocessing unit, applies rule models and dynamic thresholds to determine the smooth, slow, and congested levels of regional road segments in real time, and estimates the congestion index and passage time, generating a road condition map with timestamps.

[0047] As an option in this embodiment, the multi-level discrimination mechanism of the road condition state fusion evaluation unit includes:

[0048] Establish a dynamic rule base containing a basic traffic capacity matrix for road segments, and input standardized data into a hierarchical rule engine: the first layer determines the basic congestion level based on the ratio of real-time occupancy rate to baseline saturation flow rate;

[0049] The second layer introduces the travel time index calculated from floating car trajectory data to perform topological correlation analysis on the synchronization status of adjacent road segments;

[0050] The third layer integrates meteorological visibility data with construction event coordinates to calculate the traffic capacity reduction for areas affected by the environment; the final output includes a dynamic traffic map containing road congestion index, confidence markers, and the distribution of affected lanes.

[0051] The road condition fusion assessment unit uses a dynamic congestion index calculation formula for judgment:

[0052] ,in, Let be the fused congestion index of road segment i at time t. Real-time occupancy rate of road segments This refers to the saturation occupancy rate of road sections. The average speed of the road segment. This is the minimum speed limit for the section of road. Weather influencing factors The density of construction events is represented by α, β, γ, and δ, which are adaptive weighting coefficients. By combining exponential and logarithmic functions nonlinearly, the coupling effect between high occupancy and low vehicle speed is enhanced, transforming sudden weather and construction factors into superimposed quantitative impacts, and achieving effective aggregation of multi-source heterogeneous results.

[0053] The short-term prediction evolution unit connects the road condition map unit and the historical database unit. Combining real-time situation and historical patterns, it uses time series analysis to predict the evolution trend of traffic flow status of key road sections and road networks within the next 5-30 minute window.

[0054] Specifically, the multi-mode prediction architecture of the short-term prediction evolution unit includes:

[0055] Historical pattern matching subunit and real-time situation simulation subunit;

[0056] The historical pattern matching subunit retrieves traffic flow curves for similar scenarios from the historical database based on date type and weather characteristics; the real-time situation inference unit receives real-time congestion boundary information from the road condition map and analyzes the traffic propagation path by combining the floating car trajectory vector direction.

[0057] As an option in this embodiment, a cross-validation mechanism is initiated when a conflict is predicted: when the deviation between the real-time detected queue length growth rate and the historical pattern exceeds a threshold, the micro-inference results based on the vehicle car-following propagation model are used first to generate a future road network state evolution map with probability confidence intervals.

[0058] It should be noted that the short-term prediction evolution unit adopts an adaptive learning architecture:

[0059] Deploy a closed-loop verification module for prediction results to send the time series error values ​​between historical prediction data and actual detection data back to the pattern matching library;

[0060] Establish feature labels for special scenarios where prediction deviation exceeds a threshold, including scenarios such as sudden increases in rainfall and traffic flow after large-scale events;

[0061] When real-time environmental parameters are detected to match specific scenario labels, the system automatically switches to an emergency prediction model trained using reinforcement learning.

[0062] Furthermore, the predictive evolution calculation formula for traffic flow mutation identification in short-term predictive evolution units is as follows:

[0063] ,in, This represents the probability of sudden changes in traffic flow. Let τ be the length of the vehicle queue during a similar historical scenario. This is the queue length for real-time detection. To predict the rate of change of acceleration within a time window, λ is a road grade correction factor, with 0.8 for highways and 1.2 for urban roads. The deviation between historical patterns and real-time conditions is quantified by an exponential decay function, and the inflection point of traffic flow changes is predicted by combining the rate of change of acceleration, thereby improving the sensitivity to capturing high-volume impact events.

[0064] In addition, short-term prediction evolutionary units require adaptive training, specifically:

[0065] ,in, For the next generation of predictive evolutionary model parameter vectors, Here are the current model parameters, and η is the learning rate decay factor. Let be the gradient function of the squared prediction error. The similarity of features in the historical pattern library is used; the parameters are corrected by a sign function of the error gradient and the partial derivative of historical features, avoiding the oscillation problem of conventional gradient descent, and improving the robustness of prediction in special scenarios while ensuring the learning speed.

[0066] The intelligent route planning unit, based on the user's input of the starting point, destination, and preferences, including optimal time / distance, congestion avoidance, waypoints, real-time traffic status maps, and predicted evolution results, generates one or more congestion-avoidance optimized route planning suggestions through fusion calculation. Specifically, the congestion avoidance decision-making of the intelligent route planning unit involves:

[0067] A pre-set road network topology connectivity matrix and multi-objective cost function library are provided. After receiving user-defined cost weight parameters, the real-time traffic map is converted into a dynamic travel time weighted network.

[0068] Based on the improved heuristic search algorithm, the future multi-time road network status provided by the short-term prediction evolution unit is spatiotemporally unfolded, and the estimated travel time of each path at different times after departure is calculated.

[0069] Apply a time penalty coefficient to the path passing through the predicted congestion section to generate a set of congestion avoidance path schemes that include the range of estimated travel time fluctuations and energy consumption comparison indicators;

[0070] As an optional feature of this embodiment, the path stability calculation formula for the intelligent path planning unit is:

[0071] ,in, This is a path stability index. This is the theoretical shortest travel time. The standard deviation of historical travel time. A smoothing constant is used to prevent division by zero; the hyperbolic tangent function transforms the discrete number of congestion points and time fluctuations into a continuous stability measure, and a coefficient of 0.2 flattens out the influence of extreme values, enabling the planned path to automatically optimize between time-optimal and reliability.

[0072] The coordinated signal optimization unit receives regional congestion information output by the road condition fusion assessment unit and evolution information from the short-term prediction evolution unit, dynamically adjusts traffic signal timing schemes within the affected area, and collaboratively alleviates predicted congestion points, improving intersection traffic efficiency and road network coordination. The regional coordination mechanism of the coordinated signal optimization unit specifically involves establishing a dynamic division model for signal control sub-zones.

[0073] When the traffic condition map shows that congestion is spreading to adjacent intersection groups, the affected areas will be merged into a cooperative control domain;

[0074] Based on the platoon dispersion coefficient provided by the road condition fusion assessment unit and the queue length growth curve of the short-time prediction evolution unit, the optimal solution for the coordinated phase difference is calculated:

[0075] Provide regional traffic capacity optimization solutions, including: extending the green light time window for key flows, compressing the duration of non-congested phases, dynamically inserting all-red clearing cycles, and sending control commands to networked traffic signals for execution via V2I communication links;

[0076] Furthermore, the special event handling strategy of the coordinated signal optimization unit is as follows: an emergency vehicle passage guarantee module is configured. When a priority passage request and real-time GPS coordinates of an ambulance or fire truck are received, a green wave passage band is dynamically inserted into the coordinated signal optimization sequence based on its driving direction and predicted arrival time. The phase locking time window of each intersection in front of the vehicle is calculated, and a signal phase switching instruction chain is generated to ensure its continuous passage.

[0077] As an optional feature of this embodiment, the phase difference optimization calculation formula of the coordinated signal optimization unit is:

[0078] ,in, Let i be the ideal phase difference from intersection i to j. Let i be the actual travel time of vehicle k from i to j. The distance between intersections Characteristic speed of the road segment Let k be the number of vehicles in the convoy. Let ξ be the reciprocal of the vehicle dispersion, and let ξ be the vehicle compression compensation coefficient. By minimizing the absolute deviation between the ideal vehicle travel time and the actual time, the coupling compensation of vehicle size and dispersion is introduced into the formula to achieve multi-vehicle green wave collaborative control.

[0079] Based on the above, the interaction between the modules is as follows: the output of the multi-source data acquisition engine is directly input to the data fusion preprocessing unit; the standard data output of the data fusion preprocessing unit drives the traffic condition fusion evaluation unit; the real-time status information of the traffic condition fusion evaluation unit is input to the short-term prediction evolution unit and the intelligent route planning unit respectively; the prediction results of the short-term prediction evolution unit are also input to the intelligent route planning unit; the congestion situation information of the traffic condition fusion evaluation unit and the short-term prediction evolution unit are simultaneously input to the coordinated signal optimization unit; the control signals of the coordinated signal optimization unit act on the urban traffic signal network; the induced routes generated by the intelligent route planning unit can be fed back to the overall road network status evaluation system; among them, the database unit provides historical analysis benchmark data support for the short-term prediction evolution unit and the traffic condition fusion evaluation unit.

[0080] In summary, the present invention has the following effects:

[0081] Heterogeneous data fusion and dynamic rule engine coupling are used to improve the accuracy of road condition assessment and are compatible with abnormal scene interference: the multi-source data acquisition engine is coupled with environmental data (visibility / road surface temperature) in real time, and the signal noise interference is suppressed by the credibility weighting mechanism in the data fusion preprocessing unit; the road condition fusion assessment unit performs capacity reduction calculation on environmental data based on the road segment capacity matrix and dynamically corrects the influence weight value of meteorological visibility on traffic flow.

[0082] The mechanism solves the problem of traditional assessment systems relying too heavily on historical data under extreme weather conditions such as rainstorms / fog, enabling the congestion index calculation to adapt to environmental changes in real time, thereby ensuring that the traffic control system adapts to environmental changes during use.

[0083] A multi-mode prediction cross-validation mechanism is used to improve the robustness of short-term predictions and suppress the problem of historical data bias accumulation: the historical pattern matching subunit of short-term prediction evolution automatically retrieves historical flow curves under similar meteorological conditions, and the real-time situation inference subunit analyzes the vehicle vector propagation direction based on the road condition map; the dual-path data input cross-validation module prioritizes micro-inference based on the vehicle following principle when the flow propagation rate analyzed by floating car trajectory deviates from the historical pattern.

[0084] This mechanism effectively corrects the prediction drift problem of the traffic control system under the interference of sudden events during peak hours, and improves the narrowing of the 5-minute prediction confidence interval.

[0085] Capacity-signal control coordinated optimization is used to curb the congestion spread effect and achieve regional traffic load balance: After receiving regional congestion information from the road condition map, the coordinated signal optimization unit activates the dynamic coordinated control domain division and aggregates the green ratio control rights of affected intersection groups.

[0086] The fleet compression compensation coefficient ξ in the phase difference optimization solution calculation function is dynamically updated based on the fleet dispersion coefficient feedback value, and green light duration is extended for key flow directions to compress non-congested phases. This linkage shortens the dissipation time of congested areas and improves road network traffic efficiency.

[0087] Path stability evaluation metrics are used to enhance the robustness of planning schemes and balance time cost and reliability: the intelligent path planning unit generates multiple alternative paths based on short-term traffic flow predictions and evolution results. The path stability formula Sp undergoes a nonlinear transformation using the hyperbolic tangent function, incorporating the standard deviation of travel time. The predicted number of congestion points Ps is uniformly mapped to the interval [0,1].

[0088] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A travel traffic condition assessment and planning system based on fusion processing, characterized in that: include: Multi-source data acquisition engine, data fusion preprocessing unit, road condition fusion assessment unit, short-term prediction evolution unit, intelligent route planning unit, and coordinated signal optimization unit; The road condition fusion assessment unit, based on the fused data output by the data fusion preprocessing unit, applies rule models and dynamic thresholds to determine the smooth, slow, and congested levels of regional road segments in real time, and estimates the congestion index and transit time, generating a road condition map with timestamps. The short-term prediction evolution unit connects the road condition map unit and the historical database unit, and combines real-time situation and historical patterns to use time series analysis to predict the traffic flow status evolution trend of key road segments and road networks within a specified future time window. The output of the multi-source data acquisition engine is directly input to the data fusion preprocessing unit. The standard data output of the data fusion preprocessing unit drives the traffic condition fusion evaluation unit. The real-time status information of the traffic condition fusion evaluation unit is input to the short-term prediction evolution unit and the intelligent path planning unit, respectively. The prediction results of the short-term prediction evolution unit are also input to the intelligent path planning unit. The congestion situation information of the traffic condition fusion evaluation unit and the short-term prediction evolution unit are simultaneously input to the coordination signal optimization unit. The control signal of the coordination signal optimization unit acts on the urban traffic signal network. The induced path generated by the intelligent path planning unit can be fed back to the overall road network status evaluation system.

2. The travel traffic condition assessment and planning system based on fusion processing according to claim 1, characterized in that: The multi-level discrimination mechanism of the road condition fusion evaluation unit establishes a dynamic rule base containing the basic traffic capacity matrix of road segments, and inputs standardized data into the hierarchical rule engine: The first level determines the basic congestion level based on the ratio of real-time occupancy rate to the baseline saturation flow rate. The second layer introduces the travel time index calculated from floating car trajectory data to perform topological correlation analysis on the synchronization status of adjacent road segments; The third layer integrates meteorological visibility data with construction event coordinates to calculate the traffic capacity reduction for areas affected by the environment; the final output includes a dynamic traffic map containing road congestion index, confidence markers, and the distribution of affected lanes.

3. The travel traffic condition assessment and planning system based on fusion processing according to claim 1, characterized in that, The multi-mode prediction architecture of the short-term prediction evolution unit includes a historical pattern matching subunit and a real-time situational inference subunit. The historical pattern matching subunit retrieves traffic flow curves for similar scenarios from the historical database based on date type and weather characteristics; the real-time situation inference unit receives real-time congestion boundary information from the road condition map and analyzes the traffic propagation path by combining the floating car trajectory vector direction. A cross-validation mechanism is activated when predicting conflicts: when the deviation between the real-time detected queue length growth rate and the historical pattern exceeds a threshold, the micro-inference results based on the vehicle car-following propagation model are used to generate a future road network state evolution map with probability confidence intervals.

4. The travel traffic condition assessment and planning system based on fusion processing according to claim 1, characterized in that, The intelligent path planning unit's congestion avoidance decision model has a pre-set road network topology connectivity matrix and a multi-objective cost function library. After receiving user-defined cost weight parameters, it converts the real-time traffic map into a dynamic travel time weighted network. Based on the improved heuristic search algorithm, the future multi-time period road network status provided by the short-term prediction evolution unit is spatiotemporally unfolded, and the estimated travel time of each path at different times after departure is calculated. Apply a time penalty coefficient to paths passing through predicted congestion zones to generate a set of congestion avoidance route schemes that include the range of estimated travel time fluctuations and energy consumption comparison indicators.

5. The travel traffic condition assessment and planning system based on fusion processing according to claim 3, characterized in that, The adaptive learning architecture of the short-term prediction evolution single: Deploy a closed-loop verification module for prediction results to send the time series error values ​​between historical prediction data and actual detection data back to the pattern matching library; Establish feature labels for special scenarios where prediction deviation exceeds a threshold, including scenarios such as sudden increases in rainfall and traffic flow after large events; When real-time environmental parameters are detected to match specific scenario labels, the system automatically switches to an emergency prediction model trained using reinforcement learning.

6. The travel traffic condition assessment and planning system based on fusion processing according to claim 4, characterized in that, The intelligent route planning unit is configured with a route coordination interface for group users to receive batch route requests from the logistics dispatch center. Establish a shared road network load status matrix and mark the occupied path capacity in real time during the planning process; when multiple planned paths are detected to overlap in key bottleneck sections, staggered scheduling is carried out based on the prediction through time windows to generate a set of collaborative path schemes containing suggested departure time offsets.

7. The travel traffic condition assessment and planning system based on fusion processing according to claim 2, characterized in that, The road condition fusion assessment unit uses a dynamic congestion index calculation formula for judgment: ,in, Let be the fused congestion index of road segment i at time t. Real-time occupancy rate of road segments This refers to the saturation occupancy rate of road sections. The average speed of vehicles on the road segment. This is the minimum speed limit for the section of road. Weather influencing factors The density of construction events is represented by α, β, γ, and δ, which are adaptive weighting coefficients. By combining exponential and logarithmic functions nonlinearly, the coupling effect of high occupancy and low vehicle speed is enhanced, transforming sudden weather and construction factors into superimposed quantitative impacts, and achieving effective aggregation of multi-source heterogeneous results.

8. The travel traffic condition assessment and planning system based on fusion processing according to claim 3, characterized in that, The predictive evolution calculation formula for traffic flow mutation identification in short-term predictive evolutionary units is as follows: ,in, This represents the probability of sudden changes in traffic flow. Let τ be the length of the vehicle queue during a similar historical scenario. This is the queue length for real-time detection. To predict the rate of change of acceleration within a time window, λ is a road grade correction factor, with 0.8 for highways and 1.2 for urban roads. The deviation between historical patterns and real-time conditions is quantified by an exponential decay function, and the inflection point of traffic flow changes is predicted by combining the rate of change of acceleration, thereby improving the sensitivity to capturing high-volume impact events.

9. The travel traffic condition assessment and planning system based on fusion processing according to claim 1, characterized in that, The multi-source data acquisition engine accesses real-time precipitation intensity data provided by traffic meteorological radar stations, construction control information released by road maintenance departments, and millimeter-wave radar traffic monitoring devices deployed at urban expressway ramps. The engine has a built-in data quality verification submodule that marks abnormal sampling frequencies of floating car GPS data or broken frames in camera video streams, and activates backup sensor nodes to supplement data acquisition. Environmental data acquisition is extended to road surface temperature and visibility indicators, and it is linked with the traffic incident alarm platform to capture information on traffic accidents and sudden control events in real time.

Images