Parameter-Variable Backstepping Control for a Mobile Robot with Two Time-Scale Filter
Аннотация
Abstract Tracking the desired trajectory for mobile robots has been an important research topic. Trajectories are of different shapes and all the tracking of them are affected by the noises and disturbances. The robots should track such trajectories rapidly, but with no velocity jumps. In this paper, we used unscented Kalman filter (UKF) to reduce noises and proposed a parameter-variable backstepping controller (PVBC) with two time-scale (TTS) filter to avoid velocity jumps and reject disturbances. The unscented Kalman filter (UKF) effectively estimated postures of a mobile robot in noisy environments. The conventional backstepping controller (CBC), due to parameters fixed in no relation to its trajectory, had low responsiveness in transient-time tracking in case of rapid changes in the trajectory and often led to velocity jumps. The PVBC that can vary parameters along the trajectory improved the responsiveness of trajectory tracking and avoided velocity jumps compared to CBC. Combining PVBC and the TTS filter enabled more accurate tracking and disturbance rejection. The convergence of tracking error was guaranteed by Lyapunov stability analysis. Results showed that tracking trajectories for a mobile robot by using PVBC and TTS filter with UKF had improved both transient and steady-state performance in noisy environments.