Estimation of Road Adhesion Coefficient Using Strong Tracking CKF with M -Estimation Theory

Real-time acquisition of road adhesion coefficient (RAC) makes great sense to advanced driver assistance system. RAC cannot be obtained straightly from vehicular sensors considering the limited cost. Few studies concentrate on the estimation algorithms under model mismatch and unknown non-Gaussian noise. To deal with the challenge, a strong tracking cubature Kalman filtering on the basis of M- Estimation theory (STCKF-ME) has been proposed. Strong tracking filtering is employed to refresh the filter matrix. M- Estimation theory is adopted to guarantee the estimation precision under unknown non-Gaussian observation noise. Results demonstrate that STCKF-ME has an outstanding performance for RAC estimation when compared with CKF and STCKF. Moreover, STCKF-ME shows strong anti- disturbance ability to different driving cycles and road surface.