Bayesian calibration for Monte Carlo localization
Armita Kaboli · 2007
Localization is a fundamental challenge for autonomous robotics. Although accurate and efficient techniques now ex-ist for solving this problem, they require explicit probabilistic models of the robot’s motion and sensors. These models are usually obtained from time-consuming and error-prone mea-surement or tedious manual tuning. In this paper we exam-ine automatic calibration of sensor and motion models from a Bayesian perspective. We introduce an efficient MCMC procedure for sampling from the posterior distribution of the model parameters. We also present a novel extension of par-ticle filters to make use of our posterior parameter samples. Finally, we demonstrate our approach both in simulation and on a physical robot. Our results demonstrate effective infer-ence of model parameters as well as a paradoxical result that using posterior parameter samples can produce more accurate position estimates than the true parameters.