An Adaptive Traffic Signal Control Method Based On Microscopic Simulation

To mitigate the congestion at the intersection,we need to improve the traffic signal control strategy.Adaptive Traffic Signal Control(ATSC),as an important component of intelligent transportation system,is able to reduce vehicle delay by predicting the traffic flow and simultaneously optimizing the signal timings.In order to enhance the reliability of ATSC,new mechanisms in the traditional traffic signal control methods were proposed and established based on different data collection infrastructures and traffic arrival patterns.In the thesis,we validated effectiveness of the developed models using a microscopic simulation software,VISSIM with Component Object Model(COM)interface.The specific works in this thesis are listed as follows:For the stochastic traffic arrival pattern at isolated intersection,we proposed an improved ATSC model,comprised of a vehicle arrival estimation model and a dynamic programming algorithm.A microscopic model was developed to capture vehicle arrival dynamics and estimate the arrival time in different queue conditions.Then,NEMA phase structure was inserted into real-time traffic signal control algorithm.Three objective functions were considered,including minimization of delay,queue length and maximization of throughput.The simulation tests showed that,compared with optimal fixed signal timings,the proposed method reduced delay by around 10-20%.Minimizing delay and maximizing throughput had similar optimization performance,while minimizing queue length showed better adaption under saturated traffic demand.For the traffic arriving in platoon pattern,we proposed a platoon-based adaptive control(PASC)strategy to provide multi-modal signal control based on the on-line connected vehicle(CV)data.The key headway was introduced to identify the platoon on the road,so that a Mixed Integer Linear Programming(MILP)model was developed to optimize signal timings.The model includes NEMA phase,platoon delay constraints and implicit coordination.The results indicated that PASC model reduced 10% and 40% passenger delay for automobiles and bus respectively compared with fixed signal timings.Moreover,the fluctuation of the number of bus passengers had little influence on the control performance,and the required minimal CV penetration rate was around 20-40%.In this thesis,the traffic dynamics at intersection stop line were revealed at the microscopic level,which was helpful for reducing the vehicle delay at the intersection.The proposed model applied new traffic information collection system,which aimed to encourage its application in the future.The signal control zones ranged from isolated intersection to arterial,effectively boosting the mobility of the network.Generally,the implementation of proposed model will be beneficial for the sustainable and environment-friendly development of urban transportation system.