Virtual robot experiments for navigation in structured environments

摘要

A cost-effective and easy-to-implement mobile robot navigation scheme in a virtual robot platform is proposed. The proposed scheme consists of two cascaded phases. In the first phase, the robot is supposed to estimate its pose (position and orientation) and build a map of its operating environment using a set of radio frequency (RF) sensors. In the second phase, the robot navigates through a set of 2D points (called herein waypoints) on the ground (structured navigation). The problem in the first phase is solved using the Extended Kalman Filter-Simultaneous Localization and Mapping (EKF-SLAM) technique. The EKF-SLAM is then paired with a motion controller for the robot to navigate through waypoints. The performance of the proposed navigation scheme is evaluated using a commercial robot simulator software, V-REP (Virtual Robot Experimentation Platform), which follows industry standard robotic platforms to faithfully reproduce the proposed navigation scheme as if this navigation scheme is applied in real-life application scenarios.