CF-IDC: A robust robot’s self-localization in dynamic environments using curvature information

This paper describes a complete laser-based approach for self-localization of mobile robots. The presented algorithm is a novel variant of the classic Iterative Dual Correspondence Method (IDC). Our scan matching method improves results of this popular algorithm in dynamic environments, due to the changed elements between two consecutive scans acquired by the robot are detected and removed of the matching process. To do that, the curvature information of each element of the environment is used. The proposed scan matching algorithm consists of three stages. Firstly, the whole raw data laser is segmented into groups of consecutive range readings using a distance-based criterion and the adaptive curvature function for each group is computed. Then, this set of curvature functions is matched to the set of curvature functions associated to the previously acquired laser scan. Finally, IDC algorithm is applied only considering those scan points which belong to the set of matched curvature functions. Thus, the system is outstanding in terms of robustly, accuracy and computation time. The implemented algorithm is evaluated and compared to the classic IDC scan matching approaches. Experimental results show that the new variant of the popular IDC algorithm performs well as it adjusts in a good way to changing environments.