Design and evaluation of braced touch for touchscreen input stabilisation

Abstract

A B S T R A C TIncorporating touchscreen interaction into cockpit flight systems offers several potential advantages to aircraft manufacturers, airlines, and pilots. However, vibration and turbulence are challenges to reliable interaction. We examine the design space for braced touch interaction, which allows users to mechanically stabilise selections by bracing multiple fingers on the touchscreen before completing selection. Our goal is to enable fast and accurate target selection during high levels of vibration, without impeding interaction performance when vibration is absent. Three variant methods of braced touch are evaluated, using doubletap, dwell, or a force threshold in combination with heuristic selection criteria to discriminate intentional selection from concurrent braced contacts. We carried out an experiment to test the performance of these methods in both abstract selection tasks and more realistic flight tasks. The study results confirm that bracing improves performance during vibration, and show that doubletap was the best of the tested methods.

Brace Touch: A Dependable, Turbulence-Tolerant, Multi-touch Interaction Technique for Interactive Cockpits

“…These events enable designers to build temporal interactions (e.g., dwelling), spatial interactions [7], and spatio-temporal interactions such as gestures [27]. Other interactions rely on hardware capturing particular physical data such as pressure [12], finger contact area [23], or finger orientation [41]. Finally, finger identification consists in linking a touch event to a particular finger; this problem has proved difficult even though some research contributions are promising using geometrical analysis of touch points [5], using additional camera input [15], biometric data [37], or additional hardware worn on the fingers [28].…”

“…On fingerAdded, the reference of the created finger model is stored in place Fingers. According to Brace Touch design [12], a double tap is specified as two consecutive taps within 500 ms, a tap being a consecutive Up and Down with the same finger on the same location (spatial threshold of 30 pixels). Thus, on finger Up and Down, the finger manager produces respectively fingerPending and fingerRecovered, unless the finger does not touch the screen again, in which case a fingerRemoved event is received.…”

“…8: in vibrating conditions, braced interaction with selection using double-tap has a similar (though slightly slower) performance to non-braced interactions in non-vibrating conditions. In [12], Cockburn et al report results of an experiment examining three different forms of braced selections in turbulent and static environments. Turbulence was simulated using a Mikrolar motion platform, with a seat and touchscreen mounted on the platform.…”

“…Value of wrong target selection (not as good as false negatives) was influenced by the tasks performed by the users which were related to aircraft operations with which they were not familiar. In [12], Cokburn et al show that for the later tasks, wrong target selection errors with Brace Touch disappeared and that in 87.3% of selections, subjects elected to use braced selections during vibrations.…”

“…This paper builds on that previous work and focuses on the dependability aspects of multi-touch interactions from a holistic point of view, addressing development faults, natural faults, and operation faults (according to the taxonomy from [6]). To this end, we propose a new interaction technique called Brace Touch (introduced in [12]) where the empirical evaluation of its usability is demonstrated. This interaction technique allows flying crew to interact with touchscreens even under turbulence [11], which one of the main usability problems for touch interactions in cockpits, together with muscular fatigue and postures, as studied by Airbus human factors experts [8].…”