Gesture detection system using smart watch based motion sensors

We are about to step into a world where we would be surrounded by a large number of extremely smart wearable devices. Most of them can be considered motion capture devices in disguise owing to the sheer number and quality of sensors they possess. These devices are capable of providing data like linear acceleration, angular velocity, gravitational acceleration, compass heading, pressure and temperature with an update rate that is just enough to be interesting and useful, yet cheap and easy to acquire. We are further seeing incredible innovation in motion tracking chips present in cell phone that provide better and always-on data with minimal impact on battery life. Innovations like the Motorola X8 architecture and the Apple M7 chip are now the torch-bearers for the next era in motion tracking. It would be a waste if all this real-time data is only used by the cellphone or the smart watch that generates it. There is a whole host of amazing applications that can be conceptualized if this data is used to interact and control our computers and other devices. This paper demonstrates one such project that aims to develop ways to conveniently send and process this data to create innovative controls for computers and other common devices. However, there are various challenges we have to deal with when using such sensors. These include: high level of noise in the sensor readings, time delay in sending sensor data to the computer, detecting and ignoring unintentional actions. As our work progresses, we aim to find practical solutions to these problems.