Trajectory and slip inference on a cleaning mobile robot using simultaneous parameter and state estimation method

Abstract

This paper presents an approach to simultaneously estimating the state of a wheeled mobile robot for autonomous cleaning and parameters of dynamic friction model for describing its wheel slip. To this end, we develop a dynamic model of a wheeled mobile robot with slip, applying the LuGre dynamic friction model for describing longitudinal and lateral slip. The dynamic model is used in an unscented Kalman filter (UKF) framework. The effectiveness of the proposed method is demonstrated through simulations, whose results indicate precise estimates of the state and parameters of a wheeled mobile robot in the no-slip case.