Research And Application Of Urban Road Traffic Flow Prediction And Single Intersection Adaptive Control Based On Bayonet Big Data

With the rapid socio-economic development of domestic cities,the number of motor vehicles is growing rapidly,while the growth of urban roads is limited,which is far from keeping up with the growth of traffic demand.The problem of traffic congestion on urban roads is becoming more and more normalized and serious,and has increasingly become a "bottleneck" restricting urban development.As the smallest unit of urban traffic signal control,road intersection is the basis and key of traffic congestion mitigation.This paper takes the single intersection in the urban road network as the research object,takes the checkpoint passing data in the actual project as the AC flow data source,and studies and applies the traffic flow prediction and adaptive control of Urban Road intersection based on the existing traffic big data platform of the company.The main work and research results are as follows:1.The traffic flow prediction,artificial neural network,seq2 seq and attention mechanism are described in detail,which lays a foundation for the establishment of subsequent prediction models.According to the three mechanisms of local spatial correlation,local temporal correlation and overall spatial correlation between the urban target predicted intersection and the upstream intersection,and the time variability of the correlation mechanism,it is proposed to take the flow data of the upstream intersection as the input sequence,consider the spatio-temporal attention mechanism,and construct the urban Road short-term traffic flow prediction model through the longterm and short-term artificial neural network LSTM,Four intersections of Yueyang Avenue,the main road in Yueyang City,are selected for verification.Mae and MAPE are used as measurement indicators for overall evaluation.The average value of MAPE is 7.13%,the accuracy is nearly 93%,and the prediction accuracy is high.2.A signal adaptive control method for single intersection is designed.The traffic state information is obtained in advance by predicting the traffic state of the intersection in the next 15 minutes;Then,according to the predicted traffic state,the signal period template formulated according to the historical traffic flow data is matched to obtain the signal period length and phase of the intersection;Finally,in the 15 minute control period,the intersection signal cycle and the green signal ratio of each phase are adjusted in small steps,and the intersection of Yueyang Avenue and Nanhu Avenue in Yueyang City is selected for simulation verification.The simulation results show that the signal adaptive control method of single intersection proposed in this paper is significantly better than the multi period timing control.3.The positioning,design idea,design principle,technical route,system composition,physical structure and technical process of the traffic big data platform are introduced in detail.The traffic flow prediction function and single intersection adaptive control function based on the traffic big data platform are designed,and the interface display of the design function is carried out.