LPV torque vectoring for an electric vehicle using parameter-dependent Lyapunov functions

In this paper the application of a novel controller design approach to a torque vectoring concept is presented. An electrically driven road vehicle with two independent motors is considered, for which a good tracking of longitudinal velocity and yaw rate is to be achieved. For this vehicle an LPV model is developed in linear-fractional transformation (LFT) representation. Based on this, a gain-scheduled controller is designed using parameter-dependent quadratic Lyapunov functions and parameter-dependent shaping filters. In order to avoid performance degradations due to saturation effects, a Torque and Slip Limiter (TSL) scheme in polytopic LPV form is used. Choices both of Lyapunov functions and shaping filters turn out to have a significant influence on the controller performance; this is illustrated in simulation examples. The vehicle is simulated using a 14-degrees-of-freedom vehicle model including a Pacejka tire model. The results indicate a significant performance improvement compared to previous approaches using polytopic LPV controller design with parameter-independent shaping filters.