Dynamic Ultrasonic Hybrid Localization System for Indoor Mobile Robots

Abstract

An accurate dynamic ultrasonic hybrid localization system is presented for autonomous navigation of indoor mobile robots using multiple ultrasonic distance measurements and an extended Kalman filter (EKF). The ultrasonic sensor subsystem is composed of several ultrasonic transmitters (Txs) attached to the ceiling at known positions and several ultrasonic receivers equilaterally located on the top of the mobile robot, which has a moving speed that is not negligible. An EKF-based algorithm with a state/observation vector composed of the robot pose (or the position and the orientation) is presented using odometric and ultrasonic distance measurements. A dynamic distance estimation method is proposed to track the estimates of ultrasonic distance information from available Txs of interest using both odometric information from the robot and actual ultrasonic distance measurements. This continuous dynamic distance estimation allows persistent use of the hybrid self-localization algorithm to accurately determine the pose of the robot. The experimental results with various trajectories clearly show that the proposed method is much more accurate than only the hybrid self-localization algorithm (without the dynamic distance estimation method).