Intelligent Particle-Filter based robot localization

Abstract

The problem of the localization of a robot moving inside a closed region is considered in this paper. The localization approach used is based on the Sequential Monte Carlo Methods also known as Particle Filters. In particular we present some statistical based criteria and a logic algorithm based on those criteria to evaluate when the estimation of the position of the robot inside the region stops performing as designed due to unanticipated objects inside the region. Also presented is a fuzzy logic approach based on the same algorithm which gives a continuous localization confidence output. Based on this output a sensor model localization parameter fine tuning is presented and tested in various simulation studies.

Related Papers

• Monte Carlo Localization with Mixture Proposal Distribution(2000)
• → Mobile robot global localization using particle filters(2008)26 cited
• → Monte Carlo localization for mobile robot using adaptive particle merging and splitting technique(2010)22 cited
• → Fast Monte Carlo Localization for Mobile Robot(2011)1 cited
• → Intelligent Particle-Filter based robot localization(2011)