A Positioning Algorithm Based on Improved Adaptive Unscented Kalman Filter

Abstract

Aiming at the situation that the positioning accuracy of current mobile robots relying on traditional methods is not accurate enough, taking four-Mecanum-wheel mobile robots as the research object, in this paper, an improved adaptive unscented Kalman filter algorithm (AUKF)is proposed, combined with lidar positioning data and the result of track estimation, which can avoid the local optimal solution that may be trapped by the laser positioning result, and reduce the influence of errors caused by wheel slip and other factors. At the same time, different weights are set for the data obtained by different sensors, combined with the Sage-Husa adaptive thought, the measurement noise characteristics can be updated. The simulation results show that, under the condition that the statistical characteristics of the measurement noise are unknown, AUKF has higher accuracy than the traditional UKF algorithm in positioning results and fits the real trajectory.