Design of a Vehicle Driver Model Based on Database-Driven Control Approach

Abstract

In automobile industries, motor vehicle simulators have been utilized for developing and evaluating high-quality vehicles quickly. A driver model in a simulator is required to make the velocity of a motor vehicle model follow given target velocity in order to minimize fuel consumption. However, a conventional driver model determines the opening ratio of an accelerator and the brake pedals based on if-then rules. Therefore it cannot control the complicated vehicle models' velocity to follow the target velocity. In this paper, PII^2D control is applied as a driver model in the simulator. However, the PII^2D with fixed control parameters does not adapt to the nonlinearities of a vehicle model, thus a database-driven (DD) control approach that is effective for a nonlinear system is combined. The DD controller is one of learning control systems, and the conventional method required much time to update its database in an online manner. However, the simulators have enormous calculation time, thus the conventional updating method will take much calculation time. In this paper, an offline updating method based on the fictitious reference iterative tuning is applied for the database updating in order to avoid enormous calculation time required for simulations.