CN-122024482-A - Intelligent traffic intervention system and method based on multi-source data fusion and unmanned aerial vehicle cooperation

CN122024482ACN 122024482 ACN122024482 ACN 122024482ACN-122024482-A

Abstract

The invention discloses an intelligent traffic intervention system and method based on multi-source data fusion and unmanned aerial vehicle cooperation, and belongs to intelligent traffic and low-altitude economic fusion technologies. The intervention method comprises the steps of firstly integrating data sources through real-time acquisition and fusion processing of multi-source heterogeneous traffic data, combining fusion feature vectors constructed through space-time alignment and data cleaning operation, comprehensively and truly reflecting traffic running states, providing high-quality and all-dimensional data support for subsequent traffic event diagnosis, secondly adopting a hybrid architecture diagnosis model, considering reliability and accuracy of event diagnosis, providing accurate basis for unmanned aerial vehicle dispatching, avoiding disposal resource waste, then constructing a dispatching optimization model based on event diagnosis results and real-time states of unmanned aerial vehicle resource pools, planning optimal flight paths simultaneously, achieving refined and dynamic dispatching of unmanned aerial vehicle resources, and finally innovating a site intervention mode, and improving traffic control flexibility and effectiveness.

Inventors

Han Ruizhe
WANG BAILING
LIU HONGRI
LIU YANG
DU HAIWEN
CHEN BANGYAN

Assignees

青岛哈尔滨工业大学(威海)研究院

Dates

Publication Date: 20260512
Application Date: 20260303

Claims (10)

1. An intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation is characterized by comprising the following steps: S1, multi-source heterogeneous traffic data are collected and fused in real time; Receiving real-time traffic data from a plurality of heterogeneous data sources through an application program interface, wherein the data sources at least comprise one of GPS track and speed information reported by a vehicle navigation terminal, vehicle identity and position information collected by a road side unit and video stream data collected by an intelligent camera; s2, intelligent diagnosis and classification of traffic events based on fusion feature vectors; Inputting the fusion feature vector obtained in the step S1 into a pre-trained event diagnosis model, outputting a structural diagnosis report, wherein the diagnosis model adopts a mixed architecture of a rule-based expert system and a lightweight neural network, and diagnosis logic of the diagnosis model comprises event type judgment and severity level quantification; s3, unmanned aerial vehicle task planning and dynamic scheduling based on diagnosis results; Inquiring a task-load mapping library according to the event type and the grade of the diagnosis report, determining a required load combination, combining real-time state information of an unmanned aerial vehicle resource pool, constructing a scheduling optimization model by taking the shortest global response time as a target, solving the model, distributing and adapting the unmanned aerial vehicle for the event, and generating an optimal flight path; S4, taking the unmanned aerial vehicle as the field self-adaptive intervention of the dynamic air traffic facility; After the dispatched unmanned aerial vehicle flies to the target airspace according to the scheduling instruction, a double-cooperation self-adaptive intervention task is executed, wherein the double cooperation is respectively that air dynamic traffic signals are generated and self-adaptive adjustment and microscopic induction information interaction facing the internet-connected vehicle is carried out.
2. The intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation according to claim 1, wherein the fusion feature vector X (t) is: X(t) = [v_avg(t),σ_v(t),occupancy(t),ΔQ(t),I_obj(t)] In the formula, v_avg (t) is the average speed of a target road section at the moment t, sigma_v (t) is the standard deviation of the speed of the vehicle, the occupancy rate of a lane, delta Q (t) is the change rate of the queuing length, and I_obj (t) is the coding value of a target detection result based on video.
3. The intelligent traffic intervention method based on the cooperation of multi-source data fusion and unmanned aerial vehicle according to claim 1, wherein the diagnosis model diagnosis logic comprises event type judgment and severity level quantification; The event type judgment is that if I_obj (T) = 'accident vehicle' and v_avg (T) falls in delta T to exceed threshold alpha, the traffic accident type T_accident is judged, and if only v_avg (T) < beta and the occupancy (T) is continuously high, the traffic accident type T_ congestion is judged; the serious grade quantification is carried out according to the influence index S; S = w 1 *(1-v_avg(t)/v_free) + w 2 *(ΔQ(t)/L_segment) + w 3 *N_lanes_blocked wherein v_free is free flow speed, L_segment is road section length, N_ lanes _blocked is number of affected lanes, and w 1 、w 2 、w 3 is weight coefficient; the event class is classified into L1 (slight), L2 (general), L3 (serious) according to the preset interval in which the S value falls.
4. The intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation according to claim 1, wherein the scheduling optimization model is as follows: In the formula, U is an available unmanned aerial vehicle set, T_comm_i is communication time delay sent to the unmanned aerial vehicle i by an instruction, T_flight_i is estimated flight time of the unmanned aerial vehicle i flying event position D.loc, payload_i is a load set of the unmanned aerial vehicle i, battery_i is residual electric quantity of the unmanned aerial vehicle i, and E_min is a minimum electric quantity threshold value required for executing a task.
5. The intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation according to claim 1, wherein step S4 comprises: s4.1, generating an air dynamic traffic signal and adaptively adjusting; the dispatched UAV_k generates a visual traffic signal matched with ground road conditions in the air through a carried folding LED array or a high-brightness projection module, and the signal display parameters are dynamically adjusted according to the environment to display brightness: Wherein L_base is the basic brightness, ambient_light is the Ambient Light normalization value, traffic_Density is the Traffic Density, and k 1 、k 2 is the adjustment coefficient; s4.2, microscopic induction information interaction for the network-connected vehicles; the UAV_k establishes a direct link with a specific network-connected Vehicle entering the communication range through a C-V2X or DSRC protocol, and generates and pushes customized induction information Msg for the specific Vehicle based on real-Time observation of the unmanned plane on the field traffic flow, wherein the data structure is Msg= { vehicle_ID, action, details, validity_Time }.
6. The intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation according to any one of claims 1 to 5, further comprising: S5, real-time evaluation of the intervention effect and iterative optimization of strategies; and the background system quantifies the intervention effect in real time by comparing key indexes before and after the intervention, stores the complete data tuple of the task into a historical database, and is used for updating the diagnosis model parameters in S2 and the scheduling weights in S3 to form a self-iterative optimization technology closed loop.
7. An intelligent traffic intervention system based on multi-source data fusion in cooperation with an unmanned aerial vehicle, characterized in that it is adapted to implement the intervention method according to any of claims 1-6, said intervention system comprising: the multi-source data fusion and intelligent diagnosis subsystem comprises a data acquisition interface unit, a data fusion processing unit and an event diagnosis unit; The unmanned aerial vehicle resource scheduling and task management subsystem comprises an unmanned aerial vehicle resource library, a task planning unit and an airspace management interface; The aerial dynamic intervention and interaction subsystem is deployed on the dispatched unmanned aerial vehicle and comprises a task execution control unit, an aerial signal generation unit and a V2X communication unit; the effect evaluation and model optimization subsystem comprises an effect monitoring unit and a model iteration unit.
8. The intelligent traffic intervention system based on the cooperation of multi-source data fusion and unmanned aerial vehicle, which is disclosed by claim 7, is characterized in that the data acquisition interface unit receives real-time data streams from a vehicle navigation system, a road side unit and a traffic monitoring camera, the data fusion processing unit maps multi-source heterogeneous data to a unified space-time coordinate system through a space-time alignment algorithm to construct fusion feature vectors, and the event diagnosis unit executes event type judgment and severity level quantification through a hybrid diagnosis model to output a structured diagnosis report.
9. The intelligent traffic intervention system based on multi-source data fusion and unmanned aerial vehicle cooperation according to claim 7, wherein the unmanned aerial vehicle resource library dynamically maintains state tuples of each unmanned aerial vehicle, the task planning unit queries a task-load mapping relation database according to a diagnosis report and operates a scheduling optimization model to generate a scheduling scheme, and the airspace management interface is used for submitting a flight path and an operation airspace to an urban unmanned aerial vehicle air traffic management system to carry out real-time airspace conflict verification and flight permission application; The task execution control unit controls the unmanned aerial vehicle to fly to a target airspace based on a received scheduling scheme and real-time airborne sensor data and adaptively adjusts a hovering position, the aerial signal generation unit comprises a physical folding mechanism, a high-brightness LED array and a digital light processing projection module, the aerial signal generation unit dynamically controls traffic signal content and visual parameters displayed in the air according to received instructions and input of an environment sensor, and the V2X communication unit is used for establishing a link with a networked vehicle entering a communication radius, and packaging and sending customized induction information according to a structured format.
10. The intelligent traffic intervention system based on multi-source data fusion and unmanned aerial vehicle cooperation according to claim 7, wherein the aerial signal generating unit can automatically expand the LED array or the projection screen after the unmanned aerial vehicle takes off and automatically furl before landing; When the V2X communication unit sends the guidance information to the networked vehicle, the information of the key influencing position vehicles is processed preferentially, wherein the key influencing position vehicles comprise head vehicles and tail vehicles in a crowded motorcade, large-sized freight vehicles, emergency vehicles and vehicles with abnormal driving behaviors detected.

Description

Intelligent traffic intervention system and method based on multi-source data fusion and unmanned aerial vehicle cooperation Technical Field The invention belongs to the technical field of intelligent traffic and low-altitude economic fusion, and particularly relates to an intelligent traffic intervention system and method based on multi-source data fusion and unmanned aerial vehicle cooperation. Background With the continuous increase of urban traffic flow, the handling efficiency of sudden congestion and traffic accidents becomes a key measure of urban management level. The prior art architecture has a number of breakpoints of the "sense-decision-execution" chain when dealing with such dynamic events. In the event sensing aspect, the prior art mainly relies on fixed monitoring equipment and passive alarm. The fixed camera has a visual field blind area, floating car data is not represented enough under low permeability, various data sources are mutually independent, and an effective intelligent fusion analysis means is not available, so that the diagnosis precision on the event type, severity and influence range is not enough, and response delay is obvious. At the event handling level, traditional approaches face structural limitations. The fixed signal lamp and the traffic sign can not be dynamically adjusted according to the sudden road condition, and an effective early warning effect is difficult to play behind an accident or a congestion point, which is an important cause for frequent secondary accidents. The movable temporary facilities need manual deployment, and the response time is tens of minutes, so that the quick disposal requirement cannot be met. Police resources are limited by coverage and working strength, and all-weather refined dispersion is difficult to realize. Although unmanned aerial vehicles are applied to traffic inspection, the current application mode is primary, and is mostly used as an 'aerial camera' for post-evidence collection, the task triggering, the flight control and the intelligent coordination of a traffic command system are insufficient, the function is single, and the active intervention potential of the 'intelligent mobile terminal' cannot be exerted. The structural defects of front end perception and rear end decision disconnection and command decision and dynamic execution means disconnection in the prior art can not be overcome, and an innovation system is urgently needed to open the break points, so that automatic and intelligent closed loop of traffic event handling is realized. Disclosure of Invention The invention provides an intelligent traffic intervention system and method based on multi-source data fusion and unmanned aerial vehicle cooperation aiming at the technical problems in the prior art. In order to solve the technical problems, the invention provides the following technical scheme: the invention firstly provides an intelligent traffic intervention method based on multi-source data fusion and unmanned aerial vehicle cooperation, which comprises the following steps: S1, multi-source heterogeneous traffic data are collected and fused in real time; Receiving real-time traffic data from a plurality of heterogeneous data sources through an application program interface, wherein the data sources at least comprise one of GPS track and speed information reported by a vehicle navigation terminal, vehicle identity and position information collected by a road side unit and video stream data collected by an intelligent camera; s2, intelligent diagnosis and classification of traffic events based on fusion feature vectors; Inputting the fusion feature vector obtained in the step S1 into a pre-trained event diagnosis model, and outputting a structural diagnosis report, wherein the diagnosis model adopts a mixed architecture of a rule-based expert system and a lightweight neural network, and diagnosis logic of the diagnosis model comprises event type judgment and severity level quantification; s3, unmanned aerial vehicle task planning and dynamic scheduling based on diagnosis results; Inquiring a task-load mapping library according to the event type and the grade of the diagnosis report, determining a required load combination, combining real-time state information of an unmanned aerial vehicle resource pool, constructing a scheduling optimization model by taking the shortest global response time as a target, solving the model, distributing and adapting the unmanned aerial vehicle for the event, and generating an optimal flight path; S4, taking the unmanned aerial vehicle as the field self-adaptive intervention of the dynamic air traffic facility; After the dispatched unmanned aerial vehicle flies to the target airspace according to the scheduling instruction, a double-cooperation self-adaptive intervention task is executed, wherein the double cooperation is respectively that air dynamic traffic signals are generated and self-adaptive adjustment and microscopic indu