Cost-effective active localization technique for mobile robots

Abstract

Mobile robot localization is the problem of determining the position of a mobile robot from sensor data. Active localization provides setting the robot's motion direction and determining the pointing direction of the sensors during localization so as to most efficiently localize the robot. This paper proposes an active localization approach that employs Monte Carlo Localization, which is based on particle filters. The technique offers two main advantages. 1) The framework applies a different way of initializing the particles that helps to reduce some steps of localization, and 2) a new resampling scheme is used to reduce the cost of localization and solve the kidnapped robot problem. Experimental results show that the probability of robot successfully localize itself is considerably high, i.e. robot can recover from failure and localize itself based on new sensor data and reduction of cost is noticeable.