Use Of Wireless Signal Characteristics For Mobile Robotics Localization And Access Point Mapping

Mobile robotics is an interesting and challenging emerging field in computer science.To properly achieve true autonomy, a robot must be able to determine where it is in relationshipto a global frame of reference and it must be able to do so with a minimum of interaction fromhuman operators and impose as few constraints on its surroundings as possible. The uncertainnature of mobility and perception requires that advanced probabilistic inference techniques beapplied to minimize error. This work examines the applicability of radio-frequency signals to the mobile robotlocalization problem to localize quickly over a wide area. Particle filtering techniques areemployed to adjust for anticipated errors in both the motion model and the perception model.Radio hardware designed to capture time of flight and received signal-strength indicators (RSSI)is used to infer relative distances and triangulate the most likely position of a mobile node,taking into account a priori knowledge about past poses.