3D localization and pose tracking system for an indoor Autonomous Mobile Robot

Abstract

This paper presents a new architecture that achieves 3D localization and pose tracking for an indoor Autonomous Mobile Robot (AMR). Specifically, for speeds of up to about 0.1m/s, the architecture could localize the position with an accuracy of 1.5 cm, and determine the orientation angle to within about 4 degrees error. The localization system consists of a microcontroller embedded in the AMR and five Wireless Sensor Nodes (WSN). Two ultrasonic transmitter nodes are attached to the AMR and three ultrasonic receiver nodes are fixed on a suspended bracket. Six ultrasonic time-of-flight (TOF) measurements are used to update the AMR's pose by utilizing an Extended Kalman Filter (EKF) algorithm. In order to verify the concept, two experiment prototypes were built. In the first experiment, aiming at the precision of localization, orientation angle and slope angle, the AMR moves slowly along a slope. In the second experiment, focusing on the transient performance of the system when the orientation angle varies from one quadrant to another, the AMR moves along an arc. The results proved that the new architecture provides a high performance of localization and pose tracking for an AMR in an indoor environment.