The vehicle control system described in this paper consists of a discrete controller that computes the control parameters in two major steps: Geometric dynamic planning and plan-to-action mapping. Geometric dynamic planning computes a nominal motion, or ‘gd-plan’ of the vehicle for one control interval. Plan-to-action mapping computes the control parameters so that the gd-plan is performed. This control structure allows the independent modeling of two essential control concepts: motivation and knowledge. Motivation is modeled by computing a desired geometric-dynamic plan. Knowledge is represented through plan-to-action mapping as an approximation of the plant’s inverse system function. This paper describes the development of an extremely precise plan-to-action mapper for use in a vehicle control system that is based on empirical data. This approach is majorm step towards the development of a control structure that is indeed able to adapt to any given plant, or specifically any given vehicle. The paper first describes the development of a suiteable database capturing the vehicle’s input/output behavior. This database can then be used by a curve fitting process to find a formula for the relationship between the vehicle’s; geometric-dynamic behavior and the vehicle’s control inputs. Finally, a number of simulation runs are done to demonstrate the effectiveness of the approach.

This content is only available via PDF.
You do not currently have access to this content.