A Sliding-Mode Virtual Sensor for Wheel Forces Estimation with Accuracy Enhancement via EKF

In this paper, an algorithm for the estimation of the longitudinal and lateral forces exerted at wheel level between tires and ground is presented. Starting from a modified version of the single track vehicle model, which also includes the steady-state effect of pitch and roll on the planar movement of the vehicle, the structure is designed as a cascade of two Sub-Optimal Second-Order Sliding-Mode (S-SOSM) observers, featuring an adaptive feedback that helps improving the accuracy of the estimation of the longitudinal forces. The presented approach is purposely designed so that only standard sensors, which are usually available in commercial vehicles, are exploited. In order to alleviate the high-frequency vibrations introduced by the Sliding-Mode technique, an EKF is added as a second step, which considers the output of the S-SOSM observer as a noisy measurement, hence the virtual sensor nomenclature. The method is evaluated on experimental data, displaying good performance both in terms of accuracy and chattering alleviation.