Abstract
Robot localization is the problem of determining a robot's pose based on sensory information. This problem is consider one of the fundamental issues for autonomous robotics, hence its importance. Localization systems using wireless signal strength measurements have gained popularity in recent years, probably due to the proliferation of wireless Local Area Networks using Wi-Fi. Among these systems Gaussian Processes excel due to its flexibility and ability to model a wide variety of mappings. This paper presents an analysis and comparison of two different approaches for the localization problem based on wireless signal strength measurements. Our main motivation is the use of these learned mappings for robust localization.
