Set‐theoretic localization for mobile robots with infrastructure‐based sensing

Abstract

AbstractIn this article, we propose a set‐membership based localization approach for mobile robots using infrastructure‐based sensing. Under an assumption of known uncertainties bounds of the noise in the sensor measurement and robot motion models, the proposed method computes uncertainty sets that over‐bound the robot 2D body and orientation via set‐valued motion propagation and subsequent measurement update from infrastructure‐based sensing. We establish theoretical properties and computational approaches for this set‐theoretic localization method and illustrate its application to an automated valet parking example in simulations, and to omnidirectional robot localization problems in real‐world experiments. With deteriorating uncertainties in system parameters and initialization parameters, we conduct sensitivity analysis and demonstrate that the proposed method, in comparison to the FastSLAM, has a milder performance degradation, thus is more robust against the changes in the parameters. Meanwhile, the proposed method can provide estimates with smaller standard deviation values.