An adaptive square root cubature Kalman filter based SLAM algorithm for mobile robots

Abstract

For simultaneous localization and mapping (SLAM) of mobile robots, an innovative solution is proposed, named adaptive square root cubature Kalman filter based SLAM algorithm (ASRCKF-SLAM). The main contribution of the proposed algorithm lies that: 1) Square root factors are used in the proposed ASRCKF-SLAM algorithm to improve the calculation efficiency by avoiding the time-consuming Cholesky decompositions. 2) Using the adaptive Sage-Husa estimator to solve the large estimation errors or even divergence problem caused by the time-varying or unknown noise. Simulation results obtained demonstrate that the proposed ASRCKF-SLAM algorithm is superior to the existed SLAM method in the aspect of estimation accuracy and computational efficiency.