Development of A Glove-Based Input Device

摘要

In this paper, we proposed a new wearable input device by recognizing human hand-motions, so-called a wearable keyglove, which can solve the disadvantages of the conventional input devices and can be adapted to the mobile computing environment. This system is applied to the touch-typing-input method by means of hands for user's easy understanding and improvement of input speed, and we analyzed hand-motions for touch-typing using QWERTY keyboard because it is most widely used currently. In the analyzed result, we are able to reduce the model with 21 DOF to one that is approximated by 14 DOF. Also we developed the hand-motions capture module using low price sensors for hand-motions recognition (production cost is about $15). Typists using the keyglove system can not exactly determine the key position of desired input letter because of the virtual keyboard without the physical location of a key layout. This problem caused many typing errors at long typing time. Therefore an algorithm is needed to cope with the problem and needed to reduce typing errors due to personal differences such as hand shapes and typing habits. For this reason, we proposed the time-variant prediction algorithm for the least typing error and confirmed the possibility of the keyglove system.