Experimental Shake Gesture Detection API for Apple Watch

TL;DR

This paper presents WatchShaker, an experimental API for shake gesture detection on the Apple Watch, addressing the lack of native shake recognition. It employs an accelerometer-based heuristic where a shake is detected when $\max(|a_x|, |a_y|) > \theta$ and $t_{\Delta} > \tau$, supplemented with directional classification through simple inequalities. Although formal testing was limited, the artifact earned notable community engagement and practical interest among developers. The work contributes to wearable HCI by providing a simple, usable API and illustrating a design-science workflow for artifact-based evaluation, with potential for broader adoption and future refinements.

Abstract

In this paper we present the WatchShaker project The project involves an experimental API that detects the Apple Watchs shake gesturea surprisingly absent natively feature Through a simple heuristic leveraging the Apple Watchs accelerometer data the API discerns not just the occurrence of shake gestures but also their direction enhancing the interactivity potential of the device Despite the projects simplicity and lack of formal testing it has garnered significant attention indicating a genuine interest and need within the developer community for such functionality The WatchShaker project exemplifies how a minimalistic approach can yield a practical and impactful tool in wearable technology providing a springboard for further research and development in intuitive gesture recognition

Figures (1)

• Figure 1: Diagram illustrating the WatchShaker API's directional criteria for determining shake directions.