Fuzzy Inference-Based Driver Decision Processes and Traffic Flow Simulation

摘要

Traffic simulations have traditionally used strictly deterministic, differential-equation-based, models of driving processes. Human decision making is, however, a linguistically structured rule-based process where desired outcomes are only approximately defined. The nature of human reasoning is thus better captured by fuzzy-inference logic, which this project uses to model driving behavior. individualized fuzzy inference models of different types of drivers that have been developed and incorporated into the TRAF-NETSIM microscopic traffic flow simulation environment. Users can now choose this fuzzy inference based driver decision model as an option when running TRAF-NETSIM to simulate more realistic vehicle motion accuracy.