Continuous Input for Gesture Detection from Wrist-Worn Wearables

Abstract

Smartwatches have been readily available in markets for more than 5 years. However, smartwatches today continue to use input modalities such as touch screens, which are suitable for smartphones, but provide problems when used with small screens, such as that of a watch. Since providing input from smartwatches is cumbersome, they are used as mere extensions of smartphones. This paper presents a method to gather a continuous stream of data collected by inertial sensors of a smartwatch and smartphone to recognize different gestures as input for a smartwatch. Typical smartwatch sensors will gather data that includes movement not only of the wrist, but also that of the body. By using sensors on the phone to get the orientation of the body, the wrist-worn sensors can be used to capture the movement of the wrist relative to the body without being affected by the movement of the body. The accuracy of the calculated orientation of the wrist with respect to the body is tested and the application of the system are further discussed.