새로운 휴리스틱 기반 청소 및 분리수거 작업 경로 생성 기법

초록

Existing robotic vacuums follow area-based paths without object recognition, limiting effective debris collection. This study proposes a path planning method for combined cleaning and sorting, considering both travel distance and rotation angle for efficient operation. The proposed method first identifies the type of the nearest object to the start location and uses a genetic algorithm to determine an efficient visitation sequence. Each object type is then collected and transported to its designated destination via paths computed by the A* algorithm that avoid obstacles. The fitness of a path is evaluated based on total travel distance and cumulative rotation angles, aiming to minimize both. We validated the effectiveness of the proposed method through two scenario-based simulations and real-world experiments. The results show that our method reduces path length by up to 1.78% and rotation angle by up to 5.30% compared to conventional approaches, improving overall performance.

키워드