This paper proposes an approach to analyzing and designing an intelligent vehicle controller for partially supporting the driver's operation of a vehicle. In the process of driving a vehicle, the driver has certain expectations of vehicle behavior. The driver operates the steering wheel and accelerator and brake pedals in an effort to achieve those expectations as the vehicle moves along in the driving environment. Moreover, the driver perceives the vehicle's spatial movement and determines the next driving operation based on the relationship between vehicle behavior and the expectations of that behavior. Thus, driving can be thought of as a system formed by the interaction between the driving environment, vehicle behavior and the driver's expectations of vehicle behavior. This paper proposes a model of driving environment, driver and vehicle behavior interaction as a tool for analyzing and designing a vehicle controller and vehicle control characteristics adapted to the driving environment and the driver's intentions.
This interaction model incorporates transitions representing knowledge of the driver's particular cognitive characteristics. These transitions are expressed using an extended Petri net description method and are adopted among the model rules describing driver behavior. The Hierarchical Fuzzy Integral (HFI) is used as a multipurpose decisionmaking technique that allows explicit treatment of the driving environment, vehicle behavior and the driver's intentions. The characteristic of the driver's cognition of the driving environment is treated as affordance. Based on the affordance perception between the driver and the driving environment, the differences in vehicle behavior demanded by individual drivers have been expressed in engineering terms by varying the fuzzy rules of HFI. As an example, a control procedure designed with the proposed model is applied to automatic engine braking control during downhill coasting. The simulation results show good agreement with driving test results.