Decentralized Coordination Of Connected And Automated Vehicles At Urban Intersections

With the fast development of economy and society,traffic problems such as accidents,congestion and pollution have been growing rapidly.As an essential part of the intelligent transportation system,connected and automated vehicle(CAV)has been regarded as a promising trend for future transportation.It has the potential to significantly improve traffic safety,capacity and fuel economy.This thesis focuses on two main directions: 1)to provide a brand new intersection setting with a novel traffic light control strategy for both human-driven vehicles and CAVs,and 2)to safely coordinate CAVs at signal-free intersections without stop-and-go driving under connected and automated environment.First,this thesis introduces a brand new intersection setting which combines reconfigured stop lines and variable message signs units.Vehicle-to-infrastructure and infrastructure-to-infrastructure communications are also included.A buffer waiting zone results from the new stop line so that the starting time from a standstill can be compensated and the intersection crossing time can be shortened.An improved traffic light control algorithm is adapted for the proposed stop line.Then both human-driven vehicles and CAVs can distinctly determine their strategies only from the current phase without considering the remaining phase time.Then,this thesis considers an unprotected left-turn scenario at a signal-free intersection,where unprotected left-turn maneuvers of CAVs are emphasized.A decentralized coordination mechanism is presented for CAVs to safely cross the intersections without stop-and-go driving and collisions.The objective is to minimize fuel consumption subject to maximum throughput.The geometry of lateral and rear-end collisions and Hamiltonian analysis are leveraged to solve the problem.Guidances for the design of the system parameters are provided also.Last,this thesis focuses on a two-way six-lane scenario at a signal-free intersection where all movements of CAVs are included.A decentralized mechanism is also proposed for CAVs to safely cross the intersections without collisions and stop-and-go driving,for minimal fuel consumption subject to maximal throughput.A time occupancy model is built to avoid collisions at the intersection and a fuel consumption model is built to control the trajectory of each vehicle.Dynamic programming is leveraged to solve a multiple traffic flow queuing problem for maximal throughput.Simulations validates the effectiveness of our proposed mechanism in each scenario.By comparing with the pre-timed and actuated traffic signal control algorithms,our mechanisms significantly improves the traffic throughput and fuel consumption.