Map Boundary Optimization Based on Adaptive Iterative Extended Kalman Filter

摘要

This article mainly discusses the map boundary optimization problem in the process of mobile robot constructing the map. Aiming at the nonlinear system with environmental noise during the map construction, this paper uses iterative extended Kalman filter (IEKF) to filter the map boundary, On this basis, an improved iterative extension is proposed by combining with the adaptive noise estimator, which is defined as adaptive iterative extended Kalman filter (AIEKF). Experimental results demonstrates that the proposed AIEKF can validly reduce the boundary error during map construction. Comparing the LiDAR data processed by IEKF, adaptive extended Kalman filter (AEKF), and AIEKF with the original data, respectively, the average distance root mean squared error (RMSE) of AIEKF is reduced by about 75.8%, the average distance RMSE of AEKF is reduced by about 67.1%, and the average distance RMSE of IEKF is reduced about 68.9%.