Influence of two SLAM algorithms using serpentine locomotion in a featureless environment

Abstract

Snake-like robots are capable of different locomotion patterns to move in narrow spaces, even with uneven terrain or highly constrained environments such as tunnels or pipes. Localization in such spaces is difficult, especially because of the lack of features to recognize movement or the area specifics. In these featureless environments, odometry-less robots will be similar to a new kidnapping problem. In this paper, how the influence of a Simultaneous Localization and Mapping (SLAM) algorithm is had in featureless environments without kidnapping recovery algorithms following a certain locomotion, such as the serpentine locomotion, is presented. In a comparison between two similar SLAM algorithms, how a key process, which is only present in one of SLAM algorithms, can affect the mapping performance, is shown. Furthermore, tests modifying the parameters of this process have been done. Also, simulations while changing the grid size of the map were conducted. Simulations and experiments have been done to show the validity of our analysis.