Consolidating computer operation and wheelchair control
Torsten Felzer, Rainer Nordmann · 2007 · Proceedings of the 9th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS '07) · doi:10.1145/1296843.1296891
Summary
This short ASSETS 2007 demo proposal from Torsten Felzer and Rainer Nordmann at Darmstadt University of Technology describes the integration of two previously separate assistive systems they had built for people with severe physical disabilities. The first, HaMCoS (HAnds-free Mouse COntrol System), monitors the muscular activity of a single user-chosen muscle (for example raising the eyebrow), filters the signal stream for intentional contractions, and translates the resulting events into mouse cursor movement and two-button click actions, allowing the user to fully operate a Windows PC without keyboard, mouse, or any hand use. The second, HaWCoS, applies the same intentional-contraction interface to drive an electric wheelchair, replacing the standard joystick with the same single muscle sensor; the latest version exists as both a stand-alone microcontroller-plus-display hardware unit and as a Windows-based software simulator. Because both target populations almost completely overlap (someone who needs hands-free PC control is also likely to use a powered wheelchair), the authors built a "plugin" that runs both tools as separate Windows processes sharing the single microphone-based muscle-input channel via Windows inter-process messaging, with one application active at a time and the user switching between modes by an intentional contraction.
Key findings
The demo's contribution is the consolidation itself rather than any new evaluation results. The authors argue that interpreting both PC operation and wheelchair operation as instances of closed-loop control with visual feedback (and additional environmental feedback for the wheelchair) makes a single input channel — intentional contractions of one chosen muscle — sufficient for both tasks. Architecturally they keep the two applications as independent processes because the original tools both listened to the standard microphone input and could not run simultaneously; HaMCoS was designated as the "listener" and forwards muscle events to the wheelchair process via Windows messages when that mode is active. The practical claim is that a severely disabled user with one set of hardware (a laptop, the muscle sensor, and the wheelchair) can both operate a computer and physically relocate (e.g. move to a different room) without help, and that a caregiver is needed only for initial setup. No quantitative evaluation, user study, or accuracy metrics are reported in the demo proposal itself; the underlying HaMCoS and HaWCoS systems are referenced in earlier ASSETS, CSUN, and IEEE RAT papers.
Relevance
For practitioners and researchers in motor accessibility, this paper is a useful early example of unifying assistive interfaces across different domains — desktop interaction and powered mobility — around a single input modality so that the user only has to learn and maintain one control skill. The underlying principle (intentional muscle contractions detected by a low-cost microphone sensor, classified as discrete events, and bound to mouse, scan, or vehicle commands) has analogues in current sEMG-based control systems for prosthetics, mobility devices, and AAC, and the consolidation argument remains relevant in any context where users must currently re-learn a separate control vocabulary for each piece of assistive equipment. Limitations are substantial because this is a two-page demo proposal: there is no evaluation, no comparison to alternative single-switch interfaces (sip-and-puff, eye gaze, head switches), no detail on signal processing or error rates, and the underlying systems run on Windows XP and Visual C++ which dates the implementation. Readers should treat this as a pointer to the broader Felzer body of work on intentional-contraction interfaces rather than as a self-contained study.
Tags: hands-free interaction · muscle contraction · electromyography · powered wheelchair · mouse emulation · alternative input · motor accessibility · physical disability · human-computer interaction · assistive technology · single switch