Vehicle lateral control using a robust tracking controller based on vision look ahead system

Abstract

The problem of controlling the motion of a vehicle entering a desired path with specific road curvature, is studied using a robust PI type controller. The lateral error of the vehicle from a desired path is measured using a look-ahead vision system. Using measurements from the vision system and the vehicle variables a robust controller is computed via a recursive algorithm based on the results of Hurwitz invariability combined with simulated annealing algorithms. The effectiveness of the robust controller is illustrated through simulations for an unknown road curvature (unknown step type disturbance) and for a wide range of physical uncertainties (longitudinal velocity, vehicle mass, vision systems look ahead distance and wheels cornering stiffness). Furthermore, using a X-Y graph the vehicle path following for an unknown road profile is illustrated.