Narrow Results

This study presents a novel mouth-controlled text input (McTin) device that enables users with severe disabilities to access the keyboard and mouse facilities of a standard personal computer via the input of suitable Morse codes processed by sliding window averaging and a fuzzy recognition algorithm. The device offers users the choice of four different modes of operation, namely keyboard-, mouse-, practice-, and remote-control mode. In the keyboard-mode, the user employs a simple mouth-controlled switch to input Morse codes, which the McTin device then translates into the corresponding keyboard character, symbol, or function. In the mouse-mode, the user is able to control the direction of the mouse movement and access the various mouse functions by inputting a maximum of four Morse code elements. The remote-control-mode gives the user the ability to control some of the functions of household appliances such as TV, air conditioner, fan, and lamp. Finally, the practice-mode employs a training environment within which the user may be trained to input Morse codes accurately and quickly via the mouth-controlled switch. Although this study presents the use of a mouth-controlled switch for the input of Morse codes, the form of the input device can be modified to suit the particular requirements of users with different degrees of physical disability. The proposed device has been tested successfully by two users with severe spinal cord injuries to generate text-based articles, send e-mails, draw pictures, and browse the Internet.

The human system is a complex time varying and nonlinear system, of which pulses are the output. In order to objectively and quantitatively identify and study pulse signals, a method combining artificial neural networks (ANN) and fuzzy theory was applied to classify and recognize pulse signals. Fuzzy neural networks (NN) were adopted to perform weight adjustments according to fuzzy reasoning rules, which improved the reliability and accuracy of the pulse classification results. The fuzzy NN learning algorithm was applied to the fuzzy reasoning process according to the new fuzzy reasoning rules of data extension, realizing the interaction between systems and environment and updating knowledge and optimization models. The algorithm also established the necessary foundation for pulse diagnostic expert systems to adapt and optimize.