탐색용 군집로봇을 위한 능동감지 기반 경로생성

초록

Swarm robots, utilizing multiple robots rather than a single one, perform missions efficiently for various purposes. Despite the increase in total cost, the greatest advantages are the reduction in mission completion time and the ability to continue mission execution using the remaining robots even if some robots malfunction. However, inter-robot cooperation and task allocation, which were not considered in single-robot systems, emerge as critical factors. Path generation for swarm robots requires not only simple collision avoidance but also ensuring that multiple robots move along a globally optimal path. This paper describes a path planning method that provides optimal paths to each robot when multiple robots explore an unknown environment and measure specific physical quantities. The primary objective is to ensure that robots do not traverse the same area repeatedly and move in the direction with the highest search efficiency when selecting the next destination. To achieve this, the spatial gradient value at the current location was estimated, and movement was directed in the direction with the largest magnitude; notably, the spatial gradient value was estimated using perturbation and correlation. Simulation results using this method demonstrated efficient exploration performance compared to methods that simply move in random directions, and methods that distribute space evenly.

키워드