Robot-aided human evacuation optimal path planning for fire drill in buildings

Human evacuation during an emergency in high-rise buildings is a life-risking task and has to be performed in the least possible time. With the advancements in robotics, deploying autonomous robots in those situations could ease human efforts in evacuation processes but pose challenges in finding the optimal path to guide the people toward the exits. This paper proposes a novel path planning for the autonomous mobile robot to lead evacuees to the nearest staircase exits. The proposed algorithm is modeled by the moving-target TSP (Traveling Salesman Problem) concept for humans who need to be evacuated and uses the Genetic algorithm optimization framework to generate the optimal path in consideration of muti-staircases as exits. We evaluated the performance of the proposed algorithm by benchmarking it with traditional algorithms in a simulated environment. Also, we conducted controlled real-time experiments to validate the proposed method with the conventional methods. In both cases, the proposed algorithm demonstrated significantly better with the gain of 22.4% than other algorithms regarding distance traveled by the robot.