Usability Study On Back-of-device Interaction

In recent years, handheld devices, smart phones as the representative, have become popular in modern society, assisting people with daily affairs, such as work, study, entertainment, travel, and so on. Touch screen technology has become the mainstream interaction technology for handheld devices owing to its intuitive, natural, and fast operations. However, usability issues, e.g. limited interaction space, occlusion problem, still exist in touch screen based front-of-device interaction. Addressing these usability issues have been the key research topics in HCI community and hotspots of top HCI conferences such as CHI, UIST, MobileHCI, and so on. Back-of-device interaction makes use of the available space behind the device for interaction purpose, thereby expanding the interaction space while keeping the device in a compact size. Back-of-device interaction can address the occlusion problem when it is used as the sole interaction technique. It can assist or be assisted by front-of-device interaction to facilitate the use of a handheld device when it is used together with other front-of-device techniques. A series of groundbreaking research work by famous scholars e.g. Patrick Baudisch, Ravin Balakrishnan, Brad Myers, Daniel Wigdor, Jacob Wobbrock, have directly promoted the study on back-of-device interaction.Since usability is the core competitiveness of interactive technologies, the study on usability is one of the most important research areas in HCI. Although attention has been paid to studies on back-of-device interaction, usability studies on back-of-device interaction are still weak. From the perspective of usability, this dissertation studies the user performance of several back-of-device interactive techniques mainly through experimental methods and statistical tools. By analyzing and comparing the experimental results, several important usability questions on back-of-device interaction had been answered. The main work and contribution of this dissertation includes the following three aspects:1. A study on user performance of using vision-based back-of-device point-and-click techniquesIn real practice, there are two interaction styles for manipulating vision-based back-of-device interaction- the interaction style of moving the device and the interaction style of moving an object. Although previous studies covered various aspects of human computer interaction research, they did not answer an important usability question: Which interaction style would get a better pointing performance? This study used Fitts’ law as the theoretical foundation, and adopted a within-subject experimental design. This work first quantitatively analyzed and compared the pointing performances of completing Fitts’ reciprocal pointing task using three vision-based back-of-device point-and-click techniques. The results showed that the technique of moving the finger, belonging to the interaction style of moving an object, achieved the best pointing performance over the other two techniques. Therefore, we concluded that priority should be given to the technique of moving the finger when deploying vision-based back-of-device pointing-and-click technique on a handheld device.2. A comparison study on users’ pointing performance: front-of-device point-and-click technique vs. back-of-device point-and-click techniquesPrevious research showed that back-of-device interaction could effectively addressing hand occlusion, thus promoting user experience. So, can back-of-device interaction serve as the replacement for front touch interaction with this advantage? In this work we made pointing performance as the evaluation criteria to answer this usability question because most of the current handheld user interfaces were still GUI based. This study used Fitts’ law as the theoretical foundation, and adopted a within-subject experimental design. We quantitatively analyzed and compared the pointing performances of completing Fitts’ reciprocal pointing tasks using three pointing techniques(vision-based back-of-device technique, touch-based back-of-device technique, and front touch technique). The results showed that the front touch technique performed significantly better than the other two techniques in terms of completion time and IP(index of performance). However, in terms of error rate, the touch-based back-of-device techniques performed significantly better than the front touch technique. Because IP, calculated through effective target width, could reflect both the efficiency and accuracy of a point-and-click technique, we concluded that the comprehensive point performance of the front touch technique was significantly higher than the other two back-of-device techniques. Therefore, from the point of view of pointing performance, back-of-device techniques were not qualified for replacing the front touch technique in ordinary use scenarios, but the touch-based back-of-device technique was competent to replace the front touch technique in some special circumstances, e.g. having special requirements in pointing accuracy or mainly acquiring small controls.3. A study on back-of-device based mode switching technology and its user performanceAccording to the previous section, back-of-device interaction cannot serve as an substitute to front touch interaction in the general case due to its disadvantage in pointing efficiecy. So we explored the use of back-of-device interaction for assisting front touch technology and improving the usability of the entire interactive system. We presented a back-of-device based mode switching technique- BackAssist, and then proved its usability through experiment. Based on the proposed technique, we analyzed and compared user performance of BackAssist based technique and traditioal technique in performing copy-and-paste tasks. The results showed that BackAssist based technique significantly promoted users’ efficiency in performing copy-and-paste tasks, but was inferior to the traditional technique in terms of error rate.