Sensory fusion of UBW-TOF-based location systems for mobile robotics

Abstract

With the increasing need for mobile robots in industrial applications, real-time location systems, which is a crucial point in these applications, has attracted attention from many researchers around the world. Thus, robot location is the process of determining the robot position and orientation in its environment. Location systems using Ultra-WideBand (UWB) have been widely used in complex urban and indoor environments. Consequently, a moving UWB tag can be located by measuring the distances to fixed UWB anchors whose positions are known in advance. The difficulty of this approach remains in the fact that the measurements are not perfect. There will always be some noise in the measurements, and because of this, position determination could contain some errors that may result in decreased accuracy. In this work, the Pozyx performance, a low-cost Ultra-WideBand (UWB) Time-of-flight (TOF) technology solution, is studied and implemented on a mobile robot, through a beacon-based location scheme. In order to reduce the impact of measurement noise and system disturbances, the readings of odometry, Pozyx measures and the information of the lines of a known navigation path are fused to improve the estimated location of the mobile robot. Therefore, the goal of this integration is to improve the accuracy of location for indoor autonomous robots. Firstly, was studied the characterisation of the Pozyx measurement error among several test conditions. Then, an Extended Kalman Filter (EKF) algorithm is implemented using two heuristics that allow the release of the filter so that it converges to the correct robot pose after it has started to diverge. Consequently, the results obtained from the different location tests performed are presented and compared, to present the precision achieved and proving the several advantages of using heuristics. Overall, this work with Pozyx system showed that it is a proper and effective tool to improve the robot location in a challenging indoor environment given its good cost/accuracy trade-off.