Abstract

The work described is part of a programme aimed at the creation of the best possible virtual driver. The programme is supported by Williams F1. Previous research on the application of optimal linear preview control theory to driving road vehicles with only modest excursions from a straight-running equilibrium state is extended into the general large-lateral-motion arena. Optimal controls are found for steady-cornering trim states of an exemplary but simple car at a given speed as a function of cornering effort, up to the practical limit. The manner in which the optimal controls change as the cornering vigour changes is discussed. Simulations of the virtual driver-controlled car are shown to demonstrate the closed- loop system following lateral path demands and the advantages of employing gain-scheduled adaptive control over a fixed control scheme are demonstrated. Combined longitudinal and lateral control by optimal use of preview information will be mentioned.

Professor Sharp is a Visiting Professor in the School of Engineering at the University of Surrey. The project also involves Mark Thommyppillai, research student, and Dr Simos Evangelou (Cambridge Engineering Graduate) both of the Departments of Electrical and Electronic and Mechanical Engineering at Imperial College London.