PATH PLANNING IN SLAM

Abstract

Simultaneous Localization and Mapping (SLAM) is a technology that helps an autonomous robot to localize itself in the unknown environment creating a map and further navigating without any human interference. Within the SLAM tech- nology we are able to make detailed maps of the environment both in outdoor and indoor spaces where GPS signals are not available. Overall, SLAM is highly effi- cient technology that can help robots to properly navigate, inspect and monitor industrial environments autonomously through proper path planning. Identify- ing the shortest collision-free path and bypassing obstacles quickly is done by path-planning that applies different algorithms and their combinations. Due to the challenging environments, sensor limitations and discrepancies in localization estimation, it can be hard for robots to efficiently bypass and arrive at the desired destination. Therefore in our thesis we aim to familiarize ourselves with the path planning algorithms to fit the turtlebot. Our research focuses on individually ex- amining the pivotal path planning algorithms ( Djikstra, A* and RRT) to test it in the application of turtlebot to identify which algorithm is suitable for this robotic device. The research is done by analyzing existing literature on path- planning to further apply the knowledge in an experiment. It was found that despite the better intrinsic capabilities of A* and RRT algorithms, the Dijkstra appears to be the most advisable algorithm for the turtlebot ).