An efficient backtracking strategy for frontier method in Sensor-based Random Tree

This paper presents an efficient backtracking strategy for frontier-based exploration in Sensor-based Random Tree (SRT). An existing backtracking of Frontier-Based SRT (FB-SRT) exploration is to just move back to the previous position which a robot already passed by. But in some cases, the robot generates a long detour to move to the position that has frontier areas. The proposed algorithm aims to make the previous FB-SRT strategy more efficient by modifying its backtracking algorithm. We show comparative simulations of the proposed algorithm and existing FB-SRT algorithm to identify the merits of the proposed strategy. The simulation results confirm that the robot can save exploration time and energy while it backtracks using the algorithm presented in this paper.