Estimation of road grade and vehicle mass via model predictive control

The paper develops an active on-line estimation scheme for road grade and vehicle mass of an automotive vehicle. The scheme combines an extended Kalman filter (EKF) to generate on-line parameter estimates and a model predictive controller (MPC) to control vehicle speed trajectory so that to enhance parameter identifiability. For the latter purpose a receding horizon optimization of a cost function which penalizes a predicted variance of the parameter estimates as well as the deviation of the vehicle speed from the desired value. Uncertainties in engine torque delivery can be accounted for with this approach and probabilistic constraints on vehicle speed to maintain vehicle operation within traffic limits can be enforced. A simulation example for a typical mid-size vehicle is reported