Research On Active Control Of Intersection Traffic Based On V2X Technology

Intelligent traffic control is an indispensable or missing part of ITS.The traffic condition of the intersection determines whether the whole road network is smooth or not.The realtime traffic flow control of the intersection is the foundation of the traffic control system.At present,due to unreasonable traffic flow control at some traffic junctions in the city,road congestion is widespread.Based on the National key Experimental Project(19AKC15),on the basis of studying the traffic flow characteristics of urban traffic intersection,this paper proposes a traffic flow control algorithm with high prediction accuracy and strong real-time control for single intersection and multi-intersection respectively,and hardware-in-the-loop simulation is used to verify the traffic flow control mode.The subject research has theoretical significance and application value to alleviate the current situation of urban road congestion and improve road traffic efficiency.The main contents of this paper are as follows:(1)The traffic flow detection technology of the integration of lidar and video vehicle detection is adopted,and the hardware-in-the-loop simulation verification system is built.Based on the comparative analysis of the traditional traffic flow detection methods of single video vehicle detection,the traffic flow detection technology based on the fusion of lidar and video vehicle detection is adopted,and the comparative experimental study on the accuracy of traffic flow detection is carried out.through the field observation and counting experiment,the single video vehicle detection method and lidar detection method are compared.The comparative experiments show that the detection errors of video vehicle detection,lidar detection and Thunder integrated detection in the number of left-turn traffic are 0.56%,0.73% and 0.The errors in the detection of right turn + straight traffic flow are2.65%,1.94% and 0.34%,respectively.(2)A traffic flow control strategy algorithm at a single intersection is proposed.Based on the study of the traffic flow characteristics of a single intersection,the Kalman filter prediction model which accords with the traffic characteristics of a single intersection is selected.The time window optimization is adopted for this prediction model.According to the ratio of the number of vehicles predicted in each phase to the total number of vehicles passing through the intersection,a fuzzy control relationship is established.Through the optimization of the fuzzy control results of traffic flow,the signal-to-green ratio of each phase is controlled at the next time.In order to achieve the control output of the traffic light signal at the intersection.Sumo simulation results show that: compared with uncontrolled and traditional Kalman filtering fuzzy control strategy,the optimization of sample passage time of the proposed control strategy is 8.91% and 3.07%,and the optimization of average stop times of samples is 12.4% and 5.43%.(3)A traffic flow control strategy algorithm for multi-intersection is proposed.On the basis of studying the traffic flow characteristics of multi-intersection and considering the periodicity and correlation of multi-intersection,the long-term and short-term memory neural network prediction model with time characteristic is optimized,and the same fuzzy processing is done as that of single intersection according to the prediction results.the signalto-green ratio of each intersection at the next moment is controlled,thus the control effect of traffic light signals of multiple intersections is realized.Sumo simulation results show that:compared with non-control and traditional LSTM fuzzy control strategy,the optimization of sample travel time of the proposed control strategy is 11.77% and 4.01%,and the optimization of average stop times of samples is 5.62% and 1.93%.(4)The hardware-in-the-loop simulation verification system of traffic flow control strategy algorithm at traffic intersection is built by using V2 X technology.Based on the traffic safety principle,the actual traffic flow in the traffic flow is sampled by V2 X technology,and it is mapped to the virtual traffic flow in the simulation system,and the hardware in the loop test is carried out.Through the statistics of the time and stopping times of random vehicles passing through the intersection,the traffic flow control strategy algorithm of the intersection is verified in reverse.In this paper,the semi-open park road of China Merchants Inspection vehicle Technology Research Institute Co.,Ltd.is used as the hardware-in-the-loop simulation experiment site.The hardware-in-the-loop results show that,compared with the traditional Kalman filter predictive fuzzy control strategy,the single intersection control strategy proposed in this paper optimizes the vehicle travel time by 9.44% and 4.07%,and the average stopping times by 47.37% and 28.57%,respectively.Compared with the uncontrolled and traditional LSTM predictive fuzzy control strategy,the multi-intersection control strategy proposed in this paper optimizes the vehicle travel time by 10.17% and 4.22% and the average stop times by 21.43% and 8.33% in multi-intersection traffic scenarios.