Comparing evaluation methods for encumbrance and walking on interaction with touchscreen mobile devices

Abstract

In this paper, two walking evaluation methods were compared to evaluate the effects of encumbrance while the preferred walking speed (PWS) is controlled. Users frequently carry cumbersome objects (e.g. shopping bags) and use mobile devices at the same time which can cause interaction difficulties and erroneous input. The two methods used to control the PWS were: walking on a treadmill and walking around a predefined route on the ground while following a pacesetter. The results from our target acquisition experiment showed that for ground walking at 100% of PWS, accuracy dropped to 36% when carrying a bag in the dominant hand while accuracy reduced to 34% for holding a box under the dominant arm. We also discuss the advantages and limitations of each evaluation method when examining encumbrance and suggest treadmill walking is not the most suitable approach to use if walking speed is an important factor in future mobile studies.

The Effects of Encumbrance and Mobility on Touch-Based Gesture Interactions for Mobile Phones

“…Target accuracy for selecting 4x6mm targets dropped from 46% to 30% when comparing unencumbered walking to holding a 3kg bag in the dominant hand, for example. Later, Ng et al [23] examined encumbered tapping behaviour on a touchscreen mobile phone when walking speed and walking method (ground vs. treadmill) were varied. The results from their target selection study showed that reducing walking speed by 20% only improved accuracy by 3.6% when selecting 5x8mm targets.…”

“…A range of realworld multitasking scenarios was evaluated which included holding smaller objects than those tested by Ng et al [21,22,23], such as cups, pens and scissors. These objects required more intricate finger adjustments and hand grips during input.…”

“…To evaluate the effects of walking, we used a similar procedure to Ng et al [22,23] where an oval route (20m in length) was set up in a quiet room. Each participant's preferred walking speed during interaction (PWS&I) was recorded at the beginning of the experiment and he/she had to maintain this speed during each condition by walking sideby-side with a human pacesetter (for more details, see [23]).…”

“…Each participant's preferred walking speed during interaction (PWS&I) was recorded at the beginning of the experiment and he/she had to maintain this speed during each condition by walking sideby-side with a human pacesetter (for more details, see [23]). As explained by Ng et al, when the two physical factors are not controlled, it is difficult to interpret the results since the effects of mobility are mixed up with encumbrance.…”

“…The growth of mobile applications (apps) used in all aspects of life and the necessity to use on-screen gestures for input mean that users interact in many different ways, often on the move and even when physically encumbered (carrying everyday objects such as bags, boxes and personal gear [13,19,21]) as shown in Figure 1. Previous encumbered and walking studies [21,22,23] have extensively examined basic tapping performance on mobile phones and reported a significant drop in targeting performance. However, more biomechanically complex actions such as dragging, spreading, pinching and rotating gestures are less well studied on small touchscreen mobile devices such as smartphones and no research has investigated the effects of mobility and encumbrance on their performance.…”