Improving Input Accuracy on Smartphones for Persons who are Affected by Tremor using Motion Sensors

Abstract

Having a hand tremor often complicates interactions with touchscreens on mobile devices. Due to the uncontrollable oscillations of both hands, hitting targets can be hard, and interaction can be slow. Correcting input needs additional time and mental effort. We propose a method for automatically correcting such inputs based on motion data, gathered both with the devices' sensors and a small wearable sensor on the finger used for tapping. The development was informed by interviews with persons with tremor. Two empirical studies showed that our method, involving both smartphone and finger motion sensors without changing the user interface, allows users with tremor to select objects with up to 40 % fewer misses.

Rethinking the Dual Gaussian Distribution Model for Predicting Touch Accuracy in On-screen-start Pointing Tasks

“…Examples of improving touch accuracy include using an offset from the finger contact point [11,45,48], dragging in a specific direction to confirm an intended target among many items [2,39,59], visualizing a contact point [53,60], applying machine learning [51] or probabilistic modeling [9], and correcting hand tremor effects by using motion sensors [44]. In addition to these techniques, researchers have sought to understand why finger touch is less accurate compared with other input modalities such as a mouse cursor.…”