Set theoretic localization of fast mobile robots using an angle measurement technique

Abstract

We study the problem of localizing a mobile robot with an onboard-device making angular measurements on the location of known but indistinguishable landmarks. Novel algorithms are proposed for: 1) efficient posture initialization based on a simple linear solution; and 2) recursive posture estimation. Derived in a set theoretic framework, the algorithms cope with non-white, non-Gaussian noise and deterministic errors. Experiments with the set theoretic posture estimator demonstrate its simplicity and effectiveness in real-world applications.