Autonomous Mobile Robot Self-Referencing with Sensor Windows and Neural Networks

Abstract- When navigating an environment a mobile robot can update its position and orientation by searching known landmarks and compare predictions Urith observations. This paper presents a method of mobib-robot self-referencing where every mapped object (obstacles to the global motion planner) in the environment can be used as potential sources of position and orientation information. This approach employs the eficiency of traversability vectors (t-vectors) f o r finding in-range ‘geometric beacons and isolating sugaces visible to a sensor. Configuration-space (C-space) buffering (growing polygons to keep motion a safe distance from objects) will reduce the search time for finding in-range geometric beacons. Finally, a small multi-layered neural network is used to provide a credence value for each predicted range that can be factored in to a filter or control strategy. This approach can be generalized to any ranging sensor that samples a region (e.g. IR sensors). 1