Mobile robot localization in an unknown environment using sonar sensors and an incidence angle based sensors switching policy — Experimental results

In this work the mobile robot localization problem in an unknown environment is faced. To solve this problem, an Extended Kalman Filter, based on measurements taken from ultrasonic sensors and only on local data, with no assumption on robot´s working environment, is proposed. A new model for the ultrasonic sensors is proposed and validated through experimental tests. A switching sensors activation policy, based on the physical characteristics of ultrasonic sensors, is devised. Such policy allows power saving still achieving good estimation performance. Experimental tests, using the robot Khepera III, show the effectiveness of the proposed sensors switching policy in solving the mobile robot localization problem.