Transient Virtual Obstacles for Safe Robot Navigation in Indoor Environments

Abstract

Service robots are expected to work in real world scenarios which are dynamic in nature. The traversable and non-traversable passages of the environment might change at different times. For example, it is common for some of the passages or areas in the map to be inaccessible due to cleaning, repair works, or other reasons. On the other hand, some passages in the environment could have a high congestion of people compared to other passages. This requires a feature to alter the path of the robots to address these dynamic changes, instead of only relying on the shortest paths criterion. To this end, this paper proposes Transient Virtual Obstacles (TVO): a method to actively block desired paths by using virtual obstacles. The virtual obstacles can be placed or removed by a user anywhere in the map for programmable intervals. The robot's plan their paths considering the virtual obstacles, thereby blocking certain paths, and planning the desired trajectories through specific paths. Another major advantage of the proposed TVO algorithm is that the existing robot path planner does not needs to be modified. The proposed TVO is tested in both simulation and real environments and results are discussed highlighting the merits of safe mobile robot navigation.