Efficient Robot Path Planning in Crowded Passages
Abhijeet A. Ravankar, Ankit A. Ravankar, Yohei HOSHINO, Yukinori Kobayashi · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2019
Safe mobile robot navigation is important in human-centered environments. This paper proposes to use fixed external cameras to estimate congestion in different passages. A heat map of the passages is generated for different times. The historical data of congestion is maintained in a database, and the paths are classified into hot regions with large crowd, or cold regions with less crowd, in different times. This enables a robot to estimate the probability of crowd in different passages of the map, at a given time, and plan a path through passages with low probability of dynamic obstacles. Robot uses this information to plan congestion free paths for safe navigation. Results suggests that the proposed algorithm increases the safety of robot navigation by lowering the probability of collision in human centered environments by planning a path through less congested passages.