Abstract
High-rise buildings have become common over the decades that including for commercial purposes like offices, schools, hotels, hospitals, etc. Human evacuation during an emergency in such buildings is always seen as a liferisking task and must be performed in the least possible time. With the current advancement in the robotics field, deploying autonomous robots in those situations could ease the evacuation process by guiding the people towards the exit. So far, numerous studies with robotic platforms have been reported in the literature focusing on human evacuation from high-rise buildings in an emergency. Most such studies either used teleoperated robots or autonomous robots with known human positions to rescue them. However, none of the existing autonomous robot rescue studies considered the human movement and their positional changes over the period during the evacuation, which eventually increases the rescue time and risks human lives. In this study, we are proposing a novel emergency evacuation algorithm for an autonomous mobile robot that considers the human’s position and their moving trajectory to generate an optimal path that aids the platform to guide the humans to quickly reach the nearest exit. The proposed algorithms are based on the moving target TSP concept and use the framework of evolutionary algorithms like the Genetic Algorithm to generate the optimized sequence of predicted waypoints for the robot to intersect the human trajectories. We evaluated the performance of the proposed algorithm by benchmarking it with the traditional algorithms like greedy algorithms in a simulated environment. Also, we conducted controlled real-time experiments to validate our proposed method with the traditional methods. In both cases, our proposed algorithm demonstrated significantly better than other search algorithms in terms of distance traveled by the robot to rescue the humans.