Real-Time Pose Estimation of a Mobile Rover Using Kinect Sensor
11 Pages Posted: 15 Apr 2020
Date Written: June 2019
Abstract
The critical challenge of vision-based control of the mobile robot is to accurately estimate the position and orientation (pose) in the exploration of unknown terrains. In this paper, first a new approach of determining random er-rors in a pose measurement of Kinect scanner's 3D datasets by frame propagat-ing transformation approach is presented. Next, 3D visual information provided by Kinect scanner is used to track the optimum (minimum error) pose path of the rover through Kalman filter. The experimental results have been performed on real terrains which demonstrate the proposed approach efficiently tracks the poses from noisy 3D datasets.
Keywords: Kinect Sensor, Rover, Kalman Filter, 3D Tracking
Suggested Citation: Suggested Citation