A Parameter Estimator for a Model Based Adaptive Control Scheme for Longitudinal Control of Automated Vehicles

In order to improve the longitudinal control behavior of automated vehicles, a predictive control scheme with an adaptive vehicle state and parameter observer is proposed. The underlying nonlinear model of vehicle and powertrain dynamics makes use of the estimated torque signal which is calculated in the engine management system, as well as of vehicle speed and acceleration measurements. An Extended Kalman Filter is implemented to both estimate filtered vehicle states and the vehicle mass. Simulation results show good convergence of the parameter estimate. The contributions of this paper build the foundation to further examine the potential of improvement in fuel savings, planning accuracy and passenger comfort.