Spatiotemporal Impact Estimation Of Highway Traffic Incidents Based On Chaotic Time Series

The highway plays an important role in the transportation industry,also contributed to the prosperity and development of economy.But the traffic incident frequency is more and more high,not only in economic terms there will be a great loss of precious time will be lost due to the impact of traffic incident.Compared with other countries,China's highway traffic safety situation is not optimistic.The speed limit of the expressway is high,when the traffic accident happens and the vehicle can't move to the emergency lane in time,the normal traffic flow will be disordered,and the potential safety hazard will also be produced.Therefore,it is necessary to analyze the trend of traffic congestion caused by traffic accidents,so as to timely release traffic information to the community,and then improve the operational efficiency and emergency response capabilities.First of all,this paper classifies and analyzes the types of traffic accidents on the expressway,which can be used as the reference for the simulation design of traffic accidents.In the event of traffic accidents,the vehicle speed changes,and gradually spread to the upper reaches of the traffic incident location,that is,traffic waves.Based on the analysis of the generation and propagation mechanism of traffic wave,we can get the influence of traffic incident.Then,through the micro simulation software VISSIM,the simulation of freeway traffic incident is carried out,and the basic parameters are set up,which makes the simulation results more close to the traffic flow data in the actual traffic events.The simulation of traffic events will be run a number of times to extract the basic traffic flow information,for the determination of the chaotic characteristics of traffic flow and traffic incident time and space to determine the scope of the impact of the data base.Secondly,chaos and the characteristics of the basic concepts of chaos are introduced.Judging method of chaotic characteristics can be divided into two categories,including the determination method of qualitative and quantitative.In this paper focuses on the quantitative determination method to determine the embedding dimension,and the basic parameters of traffic flow time series and the delay time by C-C to determine the correlation dimension algorithm,G-P algorithm is adopted to calculate.Then the volume of traffic,traffic incident under the average speed and occupancy time series of chaotic characteristics are determined,and the traffic volume,average speed and share between time series chaotic correlation analysis,by judging the chaotic characteristics of traffic flow.Thirdly,in determining the traffic volume,average speed and occupancy time series are chaotic,the traffic volume,traffic incident under the average speed and occupancy time series were predicted.Through the analysis of fuzzy Takagi-Sugeno neural network prediction model of chaos in the adaptive part of the data traffic flow time series as input parameters,systematic learning training,and then to the corresponding time sequence of traffic flow parameter is forecasted.Finally,according to the traffic volume,average speed and occupancy rate time series prediction results to determine the accuracy of traffic incident space-time effect of selecting the model variables,according to the selected independent variables to construct the space-time effect of range estimation model,and estimate results,traffic wave forecasting results and simulation data collected by the event queue is compared the analysis of the model.In this paper,the VISSIM software simulation method is used to simulate the freeway traffic events,and the time series of the basic parameters of traffic flow are collected.After the calculation method to determine the chaotic characteristics,have chaotic characteristics of basic traffic incident flow conditions were determined,and the traffic flow also exists chaotic characteristics;prediction of chaotic traffic flow time series parameters.The average speed of time series is constructed one space range estimation model,the model of the simulation event time and space range estimation results than the single judge in time or space is better.