Text Entry from Power Wheelchairs: EdgeWrite for Joysticks and Touchpads
Jacob O. Wobbrock, Brad A. Myers, Htet Htet Aung, Edmund F. LoPresti · 2004 · Proceedings of the 6th International ACM SIGACCESS Conference on Computers and Accessibility (Assets 04) · doi:10.1145/1028630.1028650
Summary
This paper presents two adaptations of EdgeWrite, a gestural text entry method originally designed as a stylus-based unistroke technique for people with tremor, reconfigured for use with power wheelchair joysticks and touchpads. The research addresses a significant gap in computer access for power wheelchair users: while joysticks have been used for mouse cursor control, no integrated text entry solution existed, forcing users to rely on on-screen keyboards like WiVik that consume screen real estate, require frequent window management, and split the user's focus of attention. The authors developed C++ implementations of EdgeWrite for both an Everest & Jennings 1706-5020 power wheelchair joystick and a Synaptics touchpad, each fitted with plastic templates that provide a square boundary with physical edges and corners. EdgeWrite's recognition relies on the sequence of corner hits rather than the full path of motion, making it inherently tolerant of tremor and imprecise movement. For the joystick version, the authors addressed electronic noise by implementing a running average filter with n=12 points, and designed a 13.75mm square template that balanced speed and accuracy. The touchpad version used a 30mm square with deflated corner zones to prevent accidental diagonal strokes. The work is situated within the broader Pebbles research project investigating how handheld devices can augment desktop computing for people with motor impairments.
Key findings
In a participatory evaluation with 7 power wheelchair users (6 with Cerebral Palsy, 1 with Multiple Sclerosis), touchpad EdgeWrite achieved the highest average speed at 1.00 words per minute (WPM), joystick WiVik was second at 0.84 WPM, and joystick EdgeWrite was third at 0.77 WPM. Error rates told a different story: touchpad EdgeWrite had the lowest total error rate at 5.11%, while joystick WiVik had the highest at 29.56%, with joystick EdgeWrite in between at 25.40%. Subjective questionnaire ratings showed participants preferred touchpad EdgeWrite across all dimensions: easiest to use, easiest to learn, fastest, most accurate, most enjoyable, most comfortable, and most liked overall. The participatory design process yielded important refinements, including redesigning the letter "k" to eliminate difficult diagonal strokes, adjusting dwell time parameters for individual users, and modifying the touchpad template edges based on participant feedback. An expert comparison across three EdgeWrite implementations (PDA stylus, joystick, touchpad) showed speeds of 23.0, 12.8, and 19.1 WPM respectively, suggesting significant improvement potential with practice.
Relevance
This research demonstrates the viability of integrating text entry directly into existing power wheelchair control devices rather than requiring separate dedicated hardware, a principle the authors call "integrated control systems." For accessibility practitioners, the key takeaway is that gestural input methods designed around physical edges and corners can be highly effective for users with motor impairments because they provide tactile stability and do not require precise path control. The participatory design approach, where real wheelchair users shaped the design through iterative feedback sessions, serves as a model for inclusive technology development. The study's main limitation is its short evaluation period, which captured only initial "walk up and use" performance rather than extended learning curves. Nevertheless, the findings suggest that touchpad-based text entry is a promising direction for wheelchair-mounted computing, and the underlying EdgeWrite technique's adaptability across input devices makes it a valuable reference for designers creating accessible text entry methods.
Tags: text entry · motor accessibility · power wheelchair · joystick · touchpad · gestural input · participatory design · assistive technology · integrated control system