The Implement Principle Research Of The Acoustic Keyboard

The humanity and simpleness of the computer is demanded more and more as the development of the computer technology. Alleviating the weight and the volume of the computer will be an important direction of the development of the modern computer, so as keyboard which is an indispensable input equipment of the computer. People are alrealy exploreing and reseaching the computer's input mode and methods continually, which break the traditional keyboard input mode or structural.A new idea of acoustic keyboard is proposed to cover the disadvantage of the existing keyboard. In our idea, a piece of paper is used, which is called paper keyboard, and printed characters whose layout is the same as that of the standard PC104 keyboard as keyboard, and a microphone array is adopted to capture the sounds generated by typing the characters on the paper keyboard, and using array signal processing to get the typing sounds' location information, then changed it into character corresponding to the typing sounds, equally embedded the character into the operating system to accomplish the function of inputting information to realize the input function just like traditional computer keyboard.The implement principle of the acoustic keyboard is introduced as well as the basic structure of this system, and a few location algorithms are intrduced based on the implement idea of acoustic keyboard. First, a new time-delay estimating location algorithm based on edge-detection and a subarray structure location algorithm are proposed, and the above algorithm is developed from the basic time-delay location algorithm, and realized high location precision for the typing sounds under high sample rate. Then, another location algorithm, called delay-correlation beamforming algorithm with the performance of restraining the noise and higher location precision is researched, with which the hypo-optimal estimation problem of the time-delay estimating source location algorithm which needs two steps is solved. Afterward, a conclusion that both the algorithms' location precision is restrained by the system sample rate was given out. So, a sort of improved near-field 2D-MUSIC sound source location algorithm, which is based on the traditional MUSIC algorithm, is presented, and much higher location precision with lower sample rate is realized. But this algorithm is so complex that it is difficult for real-time implementation, furthermore, its performance is restrained by the noise mode. Therefore, a new near-field location algorithm based on higher-order cumulant was discussed. By constructed an appropriate matrix pencil to estimate the angle information of the sources directly, this arithmetic recuded the compute complex, and restrained the color noise effectively.