Proprioceptive Localization Assisted by Magnetoreception: A Minimalist Intermittent Heading Based Approach

摘要

We report a localization method that does not rely on the perception and recognition of landmarks for a robot/vehicle traveling in urban area. This is intended to be a fall back solution when everything else fails due to bad weather or other environmental challenges. The method is also very low cost and does not require intensive computation. Named as proprioceptive localization assisted by magnetoreception (PLAM), it employs a gyroscope and a compass to sense heading changes and matches the heading sequence with a preprocessed heading graph to localize the robot. We track the sensory and map uncertainties and model them in the process to formulate a sequential Bayesian estimation problem framework. Not all cases can be successful because degenerated maps may consist of rectangular grid-like streets and the robot may travel in a loop. To analyze these, we use information entropy to model map characteristics and perform both simulation and experiments to find out typical heading and information entropy requirements for localization. We have implemented our algorithm and tested it with both simulated and physical experiments. The results have confirmed that PLAM can localize vehicles on the map for nondegenerated cases.