Robust hierarchical estimator of clutch torques for a compound power-split hybrid electric vehicle

Abstract

Abstract A hierarchical estimator for a compound power-split hybrid electric vehicle is proposed herein to individually observe interacted clutch torques. It consists of reference layer and estimation layer, and combines several classical observers. The estimator utilizes less nonlinear parameters in clutch assembly by introducing torque estimations from unknown input observers into update laws, and forms a closed-loop structure to better achieve system robustness through interactive feedbacks from each other. Uncertainties of estimation model and disturbances of real system are considered and introduced into the analysis of input-to-state stability. As a result, an algorithm of gain matrix optimization is obtained in the form of linear matrix inequality. Results from platform simulations and real car tests demonstrated that it improves overall performance of steady accuracy and fast response, and obtains excellent robustness for the complicated hybrid system.