利用震動回饋改善滑動手勢用以增加多個輸入模式

Abstract

近年來，雖然偵測手勢的技術已經被充分的探索，也能做到相當 微小的辨識。但是人類的感知與運動能力卻不足以讓人類能在微小的 手勢上做操作，這大大的降低人類與機器互動的可能性。為了有效地 提升操作空間，這篇論文提出 HapTick，藉由觸覺回饋來提升一維滑 動手勢的表達性。藉由滑動路徑間所感受到的震動次數，使用者在傳 統的一次性滑動手勢中，能夠準確地知道所選擇到的目標或是模式。 為了驗證我們的想法，我們做了三個實驗。第一個實驗中，我們發 現在大於三毫米的間距下，受試者的準確率能夠大於九成五。在第 二個實驗中，我們比較了 HapTick 以及傳統的多次滑動手勢，並且詢 問使用者執行任務時的主觀感受，結果顯示，在生理需求以及整體 喜好上，HapTick 明顯勝過傳統的滑動手勢。第三個實驗中，我們將 HapTick 應用在不同的互動情境下：手臂上、物體表面上以及空中， 並且評估其準確率即完成時間。最後我們也提出了幾個互動的情境以 及應用。

While high-resolution and miniature gesture sensing technology has been widely explored, the interaction space is still limited due to the nature of low resolution human proprioceptive sense. To better utilize the control space, we introduce HapTick, a method that discretizes one-dimensional swiping gestures with prompt tactile cues. By counting the tactile stimuli on the path of swiping, the user could effectively select numeric target in one typical swipe. We first derived the effective interval between modes. The results showed that with more-than-3mm distance between ticks, the overall accuracy of 95% can be achieved. In the second study, we compared two methods for selecting a digit ranging from 1 to 10. While there’s no differences in completion time between multiple swiping selection and HapTick (3.2 sec vs 3.4 sec), HapTick outperforms in both physical demands (5 vs. 2*) and overall preference (2.41 vs. 4.41*). Lastly, we confirm the feasibility of applying HapTick to other interaction domain, e.g.on-forearm swiping, input on 2D surface and in-air gesture, in an explorative study. Several scenarios were also proposed based on our findings.