Research On Model Predictive Control Method Of Separating Trucks From Passenger Vehicles At Freeway

Due to the difference between dynamic behavior,different kinds of vehicles may have mutual interference on freeway.Road safety is influenced significantly.According to statistics,accidents caused by trucks and passenger cars leads serious casualties and property loss,generally.Setting independent lanes for passenger cars and trucks respectively could reduce traffic accidents to some extent.Currently,most management strategies are concentrated on fixed exclusive lanes,which couldn’t be adjusted when the running state of traffic flow changes.This thesis put forward a modified model predictive control(MPC)strategy used for setting dynamic exclusive lane for trucks,in order to improve traffic efficiency.At first,establishing a classification criterion for all vehicles and put them into categories that ‘passenger car’ and ‘truck’.Then propose lane division plans for different scenarios.Presenting an active control system after the comparison between some lane division control methods.In this system,research object was divided into three sections that prediction section,transition section and control section.Prediction section was implemented for predicting traffic volumes,while transition section was implemented for vehicles to make a choice that drive into which lane.just as the name implies,control section was used to applied the lane division control strategy.Then,a modified METANET model considered vehicle types was proposed for predicting traffic state based on the classic METANET traffic flow model.The parameters needed to be calibration was confirmed,and regarded prediction error on flow and speed as optimization object.Using simulated annealing algorithm calibrate the parameters.Besides,considered modified METANET model as traffic prediction model and regarded volume on transition section as predicted label for optimization.This thesis constructed a theory for searching the optimal lane division scheme,using modified BRP road impedance function as objective function.Then a lane changing switch process was presented and variable message signs(VMS)was used for publishing the lane division scheme.Last,validating the modified MPC strategy through simulation of urban mobility(SUMO)platform.Analyzing the suitable conditions in different lane numbers scenarios after the adjustment of scaling rate and traffic demand.The performance of each control strategy was compared based on real-time adjustment of lane division scheme,operation efficiency and the instantaneous traffic status.Simulation results reveal that the modified MPC strategy could not only improve both traffic volume and vehicle speed,but also decreased the traffic delay and travel time when proportion of vehicle types changes drastically,according to dynamic adjustment of lane division schemes.Compared with the static lane division scheme and feedback control method,the modified MPC could eliminate the time-lag,so it has absolute advantages in these three strategies above.