Development and validation of a low-cost mobile robotics testbed

Abstract

This paper considers the design, construction and validation of a low-cost experimental robotic testbed, which allows for the localisation and tracking of multiple robotic agents in real time. The testbed system is suitable for research and education in a range of different mobile robotic applications, for validating theoretical as well as practical research work in the field of digital control, mobile robotics, graphical programming and video tracking systems. It provides a reconfigurable floor space for mobile robotic agents to operate within, while tracking the position of multiple agents in real-time using the overhead vision system. The overall system provides a highly cost-effective solution to the topical problem of providing students with practical robotics experience within severe budget constraints. Several problems encountered in the design and development of the mobile robotic testbed and associated tracking system, such as radial lens distortion and the selection of robot identifier templates are clearly addressed. The testbed performance is quantified and several experiments involving LEGO Mindstorm NXT and Merlin System MiaBot robots are discussed.

A comparison of problem-based and didactic learning pedagogies on an electronics engineering course

“…The low-cost, scalable video-based tracking system on the testbed allows research and education in the field of mobile robots to progress from the simulation arena in a cost-effective manner. Using the video-based identification and tracking system detailed in Johnson and Hayes, 42 real-time tracking and logging of accurate positional data for multiple robotic agents is possible, providing support for a broad selection of honors undergraduate and graduate classes in robotics.…”