UCLA bioengineers have developed a smart glove that can identify American Sign Language hand forms and translate them to text. The motivation behind the work is to enable signers and non-signers to communicate and support the teaching of signing.
The glove is relatively simple containing thin stretch sensor across the knuckles as shown above, therefore a set of measurements, relating to how far the fingers or thumb are bent, can be obtained. This pattern of measurements can then be categorised as a particular word, letter or number. It is not not clear if there is an accelerometer and gyro on the back of the hands to also capture any dynamic gestures the hand makes. Currently the system has a relatively small vocabulary of 600 words and is slow compared to fluent signers, but it illustrates the principal well. The following video shows the glove in action.
In the UK signing involves both hands and gestural movements so maybe more challenging to implement.
Source: The Wire