According to a Discovery News article, researches modified a fingerless biker glove by adding 17 electrically sensitive patches over major muscle groups in the hand and lower arm.
The detected muscle movements are then wireless transmitted to a computer which make the data analysis of translating the different changes on the muscles into written words – if the program is trained to interpret them as text writing.
The current prototype setup manages a 63 percent accuracy by only five repetition/training sessions per letter. After 35 repetitions, the computers accuracy hit 97 percent.
A quite impressive result which I remember is similar to the early voice recognition programs for the computer market.
Michael Linderman, a neuroscientist and co-author of the article published in the journal Public Library of Science One (PLoS), creating a commercial glove using Bluetooth to write one-handed text messages wouldn’t be difficult.
The only concern he has is the estimated cost of such smart air writing glove which is in the area of $100.-
If this turns out this is too expensive for the texting community, there is the option to use this technology for the medical area to detect for example the early stages of the Parkinson diseases as handwriting consists of memory, knowledge, cognition and dexterity, a process that is very sensitive to the general state of the nervous system.