The development of Robotable : infrared correction techniques to enhance marker detection : a dissertation submitted in partial fulfilment of the requirements for the degree of Bachelor of Applied Computing with Honours at Lincoln University

Abstract

Robotable is an educational tool conceived by Chris Rogers of Tufts University to inspire interest in science and creativity in engineering. End users of the table interact through several different mediums and complete challenges with programmable robots in order to learn engineering concepts. Robotable was developed by The Centre for Engineering Educational Outreach (CEEO) at Tufts University in Boston, USA and Lincoln University in Canterbury, New Zealand.

The Robotable previously used a marker tracking system (developed by Pattie, 2004) for tracking LEGO® robots on the table. The tracking system operated from above the surface of the Robotable and in the visible part of the electromagnetic spectrum. This set-up had several inherent flaws. These have been addressed in this study though the use of an infrared correction model, and by positioning the camera underneath the table. The infrared correction model has been developed in Robolab™ as an extension to Pattie's (2004) original marker tracking implementation. A specific hardware set-up has also been developed to accompany the correction model, providing a significant increase in marker detection reliability. Where lighting conditions had once limited the marker tracking, they now have little influence. The improvements were so substantial that the marker size was able to be reduced by 50% of the original linear size. Furthermore, at the reduced marker size the infrared correction model performed better than the original tracking system (a 95% marker detection rate compared with a 60% marker detection rate). Hardware limitations prevented more reliable detection rates, but the system undoubtedly has more potential provided a more powerful source of infrared light is used.