Movement is one of the fundamental ways we interact with the world around us. We can configure our limbs and bodies into innumerable poses, locomote around the environment, and manipulate objects within it. Unsurprisingly, motion tracking is a fundamental input channel for almost all forms of computing. In response, literally hundreds of approaches have been proposed to capture motion, from accelerometers to Z+RGB depth cameras. The rapidly growing area of embodied and gestural interaction underlines the continuing importance of accurate, fast, and affordable motion tracking technology.

We present Lumitrack, a novel sensing technique with several unique qualities. Foremost, our approach is extremely precise: results demonstrate sub-millimeter and sub-degree translational and rotational accuracy respectively. It also offers extremely high frame rate with low latency: peak speeds over 1000 tracking frames per second (FPS) delivered ~2.5ms behind real time. Moreover, the components of our system are relatively small and low powered (<1 watt), enabling integration into mobile devices. Finally, the necessary components could be very inexpensive: under $10 parts cost in volume. These qualities make Lumitrack a powerful and intriguing addition to the toolbox of sensors used by HCI researchers and practitioners. The sensing approach can be configured and scaled for use with a range of interactive systems including tangible games, drawing applications, and augmented reality.

Research Team: Robert Xiao, Chris Harrison, Karl D. D. Willis, Scott Hudson, Ivan Poupyrev