CN-122024476-A - Intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety

CN122024476ACN 122024476 ACN122024476 ACN 122024476ACN-122024476-A

Abstract

The invention provides an unmanned plane-based collaborative traffic safety intelligent guiding method, and relates to the technical field of traffic safety intelligent guiding. The method comprises the steps of collecting multi-source data, constructing a road network dynamic risk topological graph, generating an unmanned aerial vehicle inspection path, building a three-dimensional behavior feature model of traffic participants, generating a dynamic guiding candidate scheme library by adopting a multi-target collaborative optimization strategy, generating an optimal traffic safety guiding sequence, and generating a dynamically adjusted guiding instruction according to guiding path deviation and traffic risk grade deviation. According to the invention, the limit of the traditional monitoring blind area is broken through by constructing the full-area dynamic risk perception system, the early warning of traffic risks is realized by the routing inspection path planning accurately covered by the high-risk nodes, the accident hidden danger identification accuracy is improved, the accident response time is shortened, the accident rate is effectively reduced, the road network traffic capacity is improved, and the efficient and feasible technical scheme is provided for intelligent management of urban traffic.

Inventors

GU ZHENGHAI
HUANG CHAOCHAO
WU GANG
LIU YU
Nie Fuhao
HUANG KUN

Assignees

安徽省空安信息技术有限公司

Dates

Publication Date: 20260512
Application Date: 20260204

Claims (10)

1. The intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety is characterized by comprising the following steps: Collecting traffic network basic data, historical traffic flow data and unmanned aerial vehicle sensing equipment parameters, fusing multi-source data, and constructing a road network dynamic risk topology map; According to the dynamic risk topology map, combining the unmanned aerial vehicle sensing equipment parameters to generate an unmanned aerial vehicle inspection path; acquiring real-time traffic state data and vehicle operation parameters in the unmanned aerial vehicle inspection path, identifying the driving intention of a traffic participant, quantifying risk tolerance and modeling path preference, and establishing a three-dimensional behavior feature model of the traffic participant; generating a dynamic guiding candidate scheme library by adopting a multi-objective collaborative optimization strategy based on the road network dynamic risk topology map and real-time traffic demand; Acquiring space coordinates and a road network three-dimensional model of key nodes of the candidate schemes in the dynamic guiding candidate scheme library, and generating an optimal traffic safety guiding sequence by combining a three-dimensional behavior characteristic model of a traffic participant and applying an air-space integrated fusion field construction technology and a dynamic path planning algorithm; and acquiring traffic state change data and vehicle position dynamic update data monitored by the unmanned aerial vehicle in real time, and generating a dynamically adjusted guiding instruction according to the guiding path deviation and the traffic risk level deviation between the optimal traffic safety guiding sequence and the real-time data.
2. The unmanned plane-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the traffic road network basic data comprises road network structure data, traffic signal configuration data, road class parameters and infrastructure distribution data, the historical traffic flow data comprises time period traffic flow data, average speed data, congestion duration data and accident occurrence record data, and the real-time traffic state data comprises current traffic flow, road congestion index, signal lamp state and temporary traffic control information.
3. The intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety according to claim 1, wherein the process of constructing the road network dynamic risk topology map comprises the following steps: Data cleaning is carried out on the traffic road network basic data and the historical traffic flow data, and road network node association information in a complete traffic event is extracted; Defining a road network risk association tensor, calculating first-order risk association strength of any two road network nodes, and expanding a multi-order risk association relationship, wherein the first-order risk association strength represents a weighted average value of accident occurrence probability and traffic efficiency influence coefficient between the nodes; storing the multi-order risk association relationship in a three-dimensional tensor form, and generating a topology side weight through tensor decomposition; removing connecting edges with topology edge weights lower than a preset risk intensity threshold value to obtain a road network dynamic risk topological graph, wherein nodes in the graph represent road network key positions, and edges represent risk association relations among the nodes.
4. The intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety according to claim 1, wherein the process of generating the unmanned aerial vehicle inspection path by combining the unmanned aerial vehicle sensing equipment parameters according to the dynamic risk topology map comprises the following steps: Extracting risk weights of a high-risk node set and node association edges in a road network dynamic risk topology map, and constructing a patrol constraint model by combining unmanned aerial vehicle sensing equipment parameters, wherein the unmanned aerial vehicle sensing equipment parameters comprise a sensing distance threshold value, resolution parameters, duration and a flight speed range; Establishing an inspection path optimization objective function, and taking the highest risk node coverage as an optimization objective, wherein the highest risk node coverage is the shortest inspection total mileage and the lowest energy consumption; Solving an optimization model by adopting an improved genetic algorithm to obtain an initial inspection path, and generating an unmanned aerial vehicle inspection path comprising a departure point, an inspection node, a hovering time and a flying speed by combining a real-time traffic state correction path of a road network; And monitoring the state change of the residual electric quantity and the risk nodes of the unmanned aerial vehicle in real time, and triggering path rescheduling when a preset threshold value or the new increasing rate of the high-risk nodes exceeds a preset proportion.
5. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the process of identifying the travel intention of a traffic participant comprises: constructing a driving intention feature vector based on the speed change rate, the steering angle and the following distance in the vehicle operation parameters; classifying and training the feature vectors by adopting a deep learning model, and establishing a driving intention recognition model; and carrying out iterative correction on the identification result by monitoring the dynamic change of the vehicle operation parameters in real time, and re-triggering the intention identification flow when the consistency of the continuous frame identification result is lower than a preset threshold value.
6. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the risk tolerance quantification process for traffic participants includes: based on historical traffic event data, statistics of driving behavior adjustment probabilities of traffic participants in different risk scenes; Identifying time distribution characteristics of risk behaviors of traffic participants, and constructing a tolerance function combined with risk accumulation factors; And dynamically adjusting the risk attenuation coefficient according to the real-time traffic risk level.
7. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the process of modeling the path preferences of traffic participants comprises: Extracting historical driving path data of traffic participants, and analyzing association relations between path selection and road network characteristics, trip purposes and time period characteristics; establishing a path preference evaluation model, and calculating preference scores of different paths; And dynamically updating the path preference score in combination with the real-time traffic state data to generate a dynamic path preference model of the traffic participant.
8. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the process of generating the dynamic guidance candidate library by adopting the multi-objective collaborative optimization strategy comprises the following steps: In the road network dynamic risk topology map, calculating feasibility scores of all candidate guiding schemes which can reach the destination through different association paths by taking the current position and the destination of the traffic participant as a starting point and a destination; performing multi-target collaborative correction, and detecting the second-order influence of real-time traffic flow distribution and road network risk states on the candidate guiding scheme; and monitoring the risk change rate and the traffic efficiency attenuation rate of the candidate schemes in the pool in real time, and screening the guiding schemes according to a dynamic preset feasibility threshold value to obtain a dynamic guiding candidate scheme library.
9. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the process of generating an optimal traffic safety guidance sequence comprises: Mapping the space coordinates and risk weights of key nodes of each candidate scheme in the dynamic guiding candidate scheme library into a four-dimensional space-time risk field; superposing a three-dimensional behavior feature model of a traffic participant to form a comprehensive constraint field, and establishing a space-time optimization equation to obtain a continuous guide path curve, wherein the comprehensive constraint field comprises risk constraints defined by risk tolerance, path direction constraints determined by driving intention and driving experience constraints converted by path preference; Discretizing the continuous guide path curve into a guide node sequence, and sequencing and outputting the guide node sequence according to space-time reachability as a guide instruction set to obtain an optimal traffic safety guide sequence.
10. The unmanned aerial vehicle-based collaborative traffic safety intelligent guidance method according to claim 1, wherein the process of generating dynamically adjusted guidance instructions comprises: acquiring vehicle position coordinates and traffic risk levels monitored by the unmanned aerial vehicle in real time, and establishing a guide decision entropy monitoring model according to the position deviation and traffic risk level deviation of the real-time position coordinates and the optimal guide path; triggering a dynamic re-planning condition when the guiding decision entropy exceeds a preset threshold, wherein the dynamic re-planning condition comprises recalculating the space-time risk field gradient direction based on the latest unmanned aerial vehicle monitoring data, and generating a dynamically adjusted guiding instruction, and the dynamically adjusted guiding instruction comprises a steering adjusting instruction, a vehicle speed control instruction and a path switching suggestion.

Description

Intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety Technical Field The invention relates to the technical field of traffic safety intelligent guiding, in particular to an unmanned plane-based collaborative traffic safety intelligent guiding method. Background Along with the acceleration of the urban process and the rapid increase of the conservation quantity of motor vehicles, the problems of traffic jam, accident frequency and the like become core bottlenecks for restricting urban development, and the traditional traffic safety guiding mode gradually exposes short plates with limited coverage, delayed response, insufficient individuation and the like. The current mainstream traffic guidance system is mostly dependent on fixed monitoring equipment (such as intersection cameras and coil detectors) to construct a data acquisition network, but the equipment has the defects of high installation cost and more monitoring blind areas, and is difficult to realize full-area coverage under complex road network or temporary traffic event scenes. For example, suburb junction roads, construction road segments and other areas often lack monitoring equipment, so that traffic conditions are not perceived timely, and guide instructions are disjointed from actual demands. In the prior art, a fixed path planning strategy is mostly adopted, only the shortest distance or the shortest time is used as an optimization target, the individual difference of traffic participants is not fully considered, the guiding scheme of the behavior characteristics of the driver is ignored, the execution rate is low, and dangerous behaviors such as illegal lane changing, overspeed and the like of the driver are easily caused. Meanwhile, the data processing mode of the traditional guiding system is single, the road network risk conduction rule is difficult to accurately describe, and when traffic accidents, temporary control and other emergency occur, the path re-planning is lagged, so that traffic jam diffusion is often aggravated. Disclosure of Invention The invention provides an intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety, which is used for solving the defects existing in the prior art. The invention provides an intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety, which comprises the following steps: and collecting traffic network basic data, historical traffic flow data and unmanned aerial vehicle sensing equipment parameters, fusing multi-source data, and constructing a road network dynamic risk topology map. And according to the dynamic risk topology map, combining the unmanned aerial vehicle sensing equipment parameters to generate an unmanned aerial vehicle inspection path. Real-time traffic state data and vehicle operation parameters in an unmanned aerial vehicle inspection path are acquired, driving intention identification, risk tolerance quantification and path preference modeling are carried out on traffic participants, and a three-dimensional behavior characteristic model of the traffic participants is established. And generating a dynamic guiding candidate scheme library by adopting a multi-target collaborative optimization strategy based on the road network dynamic risk topology map and the real-time traffic demand. And acquiring space coordinates and a road network three-dimensional model of key nodes of the candidate scheme in the dynamic guiding candidate scheme library, and generating an optimal traffic safety guiding sequence by combining a three-dimensional behavior characteristic model of a traffic participant and applying an air-space integrated fusion field construction technology and a dynamic path planning algorithm. And acquiring traffic state change data and vehicle position dynamic update data monitored by the unmanned aerial vehicle in real time, and generating a dynamically adjusted guiding instruction according to guiding path deviation and traffic risk grade deviation between the optimal traffic safety guiding sequence and the real-time data. According to the intelligent guiding method based on unmanned aerial vehicle cooperative traffic safety, traffic network basic data comprise road network structure data, traffic signal configuration data, road grade parameters and infrastructure distribution data. The historical traffic flow data includes time period traffic flow data, average vehicle speed data, congestion duration data, and accident occurrence record data. The real-time traffic state data includes current traffic flow, road congestion index, signal light status, and temporary traffic control information. According to the intelligent guidance method based on unmanned aerial vehicle cooperative traffic safety, the process for constructing the road network dynamic risk topology map comprises the following steps: And data cleaning is carried out on the traffic road network basic data and the hist