Adaptive Backpropagation Algorithm for Clustered Indoor Motion Planning

Abstract

Grid based navigation is the simplest navigation methodology for unmanned ground vehicles (UGV) particularly for indoors. The grid map formation, grid cell occupancy, self-localization and avoiding obstacles in derived path are major considerations in navigation. This research work elaborate a new strategy of mobile robot navigation to goal point using an Adaptive Backpropagation tree based algorithm. For a confined stopping point like a charging station for an autonomous vehicles, this work provide minimal solution to reach that point. The path exploration begin from the stop point rather than the start point where robot is located. This backpropagation strategy implemented in real time scenario for its effectiveness. The simulation and experimental results gives the scope for robotic applications such as advance to stationary parking point or charging point.