Pathfinder: A Fog Assisted Vision-Based System for Optimal Path Selection of Service Robots

Service delivery application involving robots relies on the success of the navigation process. To ensure maximum performance, the navigation process should generate optimal path avoiding collisions with dynamic obstacles. Selecting such a path by analyzing the dynamic environment condition while putting minimal overhead on robots is a challenging problem in service robot navigation. In this work, we develop a Fog assisted service robot navigation system that puts most of the computing tasks to a central Fog server. The server employs a vision-based monitoring system to locate the robots and obstacles. The optimal path is generated by analyzing the available information and the robots are instructed accordingly. We implement a test-bed system to evaluate the performance of the proposed system. The results depict that the proposed fog-based system can achieve as low as 23.81% of average distance covered, 22.05% of average service delivery time, and 22.72% of average energy consumption compared to the without fog systems. Contribution-A fog-based system for service robot navigation is proposed that uses computer vision to generate optimal obstacle free path.