An analysis of interaction in the context of wearable computers.

The focus of this thesis is on the evaluation of input modalities for generic input tasks, such inputting text and pointer based interaction. In particular, input systems that can be used within a wearable computing system are examined in terms of human-wearable computer interaction. The literature identified a lack of empirical research into the use of input devices for text input and pointing, when used as part of a wearable computing system. The research carried out within this thesis took an approach that acknowledged the movement condition of the user of a wearable system, and evaluated the wearable input devices while the participants were mobile and stationary. Each experiment was based on the user's time on task, their accuracy, and a NASA TLX assessment which provided the participant's subjective workload. The input devices assessed were 'off the shelf' systems. These were chosen as they are readily available to a wider range of users than bespoke inpu~ systems. Text based input was examined first. The text input systems evaluated were::a keyboard,; an on-screen keyboard, a handwriting recognition system, a voice 'recognition system and a wrist- keyboard (sometimes known as a wrist-worn keyboard). It was found that the most appropriate text input system to use overall, was the handwriting recognition system, (This is forther explored in the discussion of Chapters three and seven.) The text input evaluations were followed by a series of four experiments that examined pointing devices, and assessed their appropriateness as part of a wearable computing system. The devices were; an off-table mouse, a speech recognition system, a stylus and a track-pad. These were assessed in relation to the following generic pointing tasks: target acquisition, dragging and dropping, and trajectory-based interaction. Overall the stylus was found to be the most appropriate input device for use with a wearable system, when used as a pointing device. (This isforther covered in Chapters four to six.) By completing this series of experiments, evidence has been scientifically established that can support both a wearable computer designer and a wearable user's choice of input device. These choices can be made in regard to generic interface task activities such as: inputting text, target acquisition, dragging and dropping and trajectory-based interaction.