Augmenting a miniature humanoid platform with a low-cost networked computer vision framework

My first paper as primary author was recently published on IEEEXplore. I presented this paper in January for the 2018 IEEE Computing and Communication Workshop and Conference (CCWC). Below is the abstract for my paper. The final published paper can be read here: http://ieeexplore.ieee.org/document/8301766/.

Abstract: Miniature humanoids are becoming an increasingly common platform for humanoid robotics research and education. However, the prohibitively high cost of advanced platforms such as the ROBOTIS-OP2 drives many educators and small research institutions toward cheaper options such as the ROBOTIS-Mini. While these platforms have versatility in full-body motion, they often lack computational power and vision capabilities. This paper presents the augmentation of the ROBOTIS-Mini with a camera, local processor, and networked system for computer vision. This augmented platform is referred to as Mini-CV. The Mini-CV system provides an ultra low-cost solution for computer vision that reduces the need for high on-board computational power and provides an advanced framework for networked control. A study of the latency in the system is presented and compared to that of the ROBOTIS-OP2, a popular miniature humanoid that retails for more than 20× the price of our augmented system. The results demonstrate the viability of the Mini-CV as an ultra low-cost alternative to more expensive miniature humanoid platforms.