Alternative Text Entry Using Different Input Methods
Torsten Felzer, Rainer Nordmann · 2006 · Proceedings of the 8th International ACM SIGACCESS Conference on Computers and Accessibility (Assets '06) · doi:10.1145/1168987.1168991
Summary
This paper presents the HaMCoS (Hands-free Mouse Control System) framework and its text entry applications, designed for people who cannot use their hands to operate a keyboard. HaMCoS works by detecting intentional muscle contractions — specifically single (CE1) and double (CE2) contractions of a chosen muscle such as the brow — picked up by a specialised acquisition device and microphone. These two signals are mapped to a state machine that emulates a two-button mouse, allowing the user to control cursor movement through four directional states (LEFT, UP, RIGHT, DOWN) and a STOP state that triggers mouse clicks. The paper describes two text entry applications built on this framework. The MouseEditor provides a large 8x8 on-screen keyboard where users click virtual keys using the HaMCoS-controlled cursor or a manual mouse. The LURD-Writer is a novel application where characters are selected by hitting the edges of a small input square in specific sequences (Left, Up, Right, Down — hence LURD), inspired by Morse code but using four directional "bits" instead of dots and dashes. The LURD sequences are designed to resemble the Morse alphabet for ease of learning, with rules that make sequences memorable and optimised for the HaMCoS cursor movement pattern. Both applications include word completion drawing from a general dictionary of 10,000 frequent English words and a personal vocabulary that learns from the user's input.
Key findings
Experiments with a 35-year-old participant diagnosed with Friedreich's ataxia (who normally uses a standard keyboard and mouse with reduced capability) measured the time to copy a 741-character text passage. With a manual mouse, the MouseEditor achieved 20 characters per minute (CPM) without completion, rising to 14-15 CPM with general/optimum completion (slower due to cognitive overhead of scanning completion lists). The LURD-Writer with manual mouse achieved 15 CPM without completion and 12 CPM with general completion, but the gap narrowed with optimum completion. With HaMCoS (muscle-based input), the MouseEditor reached 5 CPM without completion and 6 CPM with both completion levels, while the LURD-Writer achieved 3 CPM without completion, rising to 5-6 CPM with completion. The LURD-Writer provided approximately 30% speed improvement over the MouseEditor for HaMCoS-based input across all completion conditions, because its small input area is optimised for the HaMCoS cursor repositioning behaviour. The "temporal penalty" of using HaMCoS versus a manual mouse was approximately 75%, reducible to about 25% using the LURD-Writer with optimum completion. Word completion in HaMCoS was always beneficial (unlike with manual mouse) because the cognitive cost of scanning the completion list does not add a physical cost — the cursor is already stationary while the user reads.
Relevance
This research is significant for demonstrating that people with very severe physical disabilities can achieve functional text entry using only a single controllable muscle. The HaMCoS framework's design principle — requiring extremely little physical effort (just intentional contractions of any one muscle) — makes it potentially applicable to people with conditions ranging from Friedreich's ataxia and muscular dystrophy to high-level spinal cord injuries and ALS. The insight that word completion behaves differently depending on input method is practically important: while completion can actually slow down manual mouse users (due to the cognitive load of scanning lists), it consistently speeds up HaMCoS users because scanning the list costs no additional physical effort. The LURD-Writer's approach of encoding characters through directional edge-hitting sequences, optimised for minimal cursor travel, represents a creative solution to the fundamental challenge of mapping a small number of input signals to a large character set. The paper also notes the potential to combine HaMCoS with wheelchair control devices, since both use the same type of bio-signal input.
Tags: text entry · motor impairment · alternative input · single switch · bio-signals · brain-computer interface · word completion · Morse code · hands-free · EMG