Traffic Flow Modeling and Simulation Based on A Novel Cellular Learning Automaton

A novel cellular learning automaton traffic flow model is proposed to solve the problem that the probability of randomization in the NaSch model is not consistent with the actual traffic. The learning mechanism is introduced in this model. Cellular can learn traffic information from cellular neighbors in real time. The traffic environment information, such as, relative speed and safety distance will be cellular parallel ruler through the randomization probability form after learning. Finally through numerical simulation, the space-time characteristics were obtained, and comparison with the NaSch model is analyzed. The results show that the improved model can reduce the blocking the road to a certain extent and traffic jam dissolving efficiency is higher. The stability analysis of vehicle running from two aspects of speed fluctuation and headway fluctuation shows that the method in this paper can make traffic flow more stable.