A hybrid data association approach for mobile robot SLAM

Abstract

Data association is critical for the simultaneous localization and mapping (SLAM) of mobile robots. The classic data association algorithms have their own advantages and disadvantages, such as individual compatibility nearest neighbor (ICNN) algorithm and joint compatibility branch and bound (JCBB) algorithm. In this paper, we present a hybrid approach of data association based on local maps by combining them. ICNN is firstly used to do data association in the local map whose arrange is determined by the preset threshold. In order to overcome the problem of low reliability of ICNN, the errors detection in the data association results is necessary. If there are mismatchings, JCBB will be used to correct them in the local area around mismatched measurements to enhance the correct rate. The experimental results show that the proposed method performance of the speed and accuracy is satisfactory, even in the complex environments.