Research on Merging Pattern after Toll Based on Simulation

摘要

We establish a fan-in system (including combination area design, merging pattern and so on) which can best prevent accidents, improve throughput and minimize the cost. To achieve our goal, first cellular automaton is used to stimulate the actual toll plaza and establish some kinds of combination area with different shapes and sizes so that we can estimate the cost of construction. In this process, we also use Nagel-Schreckenberg (NS) vehicle-following model and select 3 merging patterns to determine the movements of our "cars". By doing this, we get sufficient data from simulation, such as average velocity of moving cars, sharp braking frequency, throughput, etc. Finally, we establish social cost analysis model to obtain a relatively objective evaluation. We calculate the economic cost of accident per year, the time cost per year measured by money and the construction fee of the whole plaza shared by each year. Thus, we can compare different solutions and reach the optimization. The best solution is the symmetrical narrowing shape, Merging as soon as possible. To test our solution, we put it in different conditions including various traffic density, self-driving cars. We also find out the divergent influence of different kinds of tollbooth. One more interesting results is that by applying basic optimizing model and some simple simulations. we can at last find the best proportion of ECT, AT and MTC under certain circumstances.