Improved SLAM algorithm using fuzzy filter and curvature data association

The issue of simultaneous localization and mapping (SLAM) is an excellent technology. Normally, the current measurements need to be compared with all existing landmarks. However, the accuracy of the estimated location of the robot will decrease because of incorrect data association. To solve these problems, this paper presents a novel architecture for SLAM. The fuzzy filter and curvature data are used to filter current measurement to retain special measurements and avoid wrong landmarks. In addition, triangulation is used to improve the accuracy of the robot's location. The effectiveness of the proposed algorithm is showed by means of simulation results.