Research And Implementation Of Desktop-based Virtual Keyboard For Mobile Phone

Despite the increasing sophistication of mobile technology,interacting with today's mobile devices can involve painful contortions.The soft keyboard based on the portable devices display is the mainstream of human-computer interaction,but miniature circuits and liquid crystal display technology keep pushing portable devices to a smaller form factor,making the size of the soft keyboard on display also getting smaller and smaller.However,human fingers and hands do not shrink accordingly,which makes the soft keyboard-based touch-screen input inefficient and error-prone,restricting the development of modern mobile devices.In this paper,we propose the desktop-based virtual keyboard for the phone input.Users can quickly enter a variety of characters on any desktop with both hands,just like tapping on the physical keyboard.Its core idea is to use the dual microphone on the phone to get click signal when users tap the virtual keyboard,and precisely calculate the time difference of arrival,so as to accurately identify the user's click location,so that enable text-input experience similar to that on a physical keyboard.Further,we use the correlation between the keys when we enter the English word to reduce the error of the key recognition,so as to obtain the more accurate complete word input.In this paper,we test the design of the program on the actual desktop and mobile phones,the test results show that the performance of the program basically meet the expected design requirements.The contributions of this paper breaks down into the following aspects:(1)we proposed a centimeter-level sound source localization method on solid surface based on cross correlation and source signal peak detection,which can realize81% adjacent key recognition accuracy;(2)observing the particularity of tapping the space bar when the keyboard is typed,we proposed a method of distinguishing the word by identifying the space bar,and usingthe special spectral structure of the click signal when the space bar is tapped;(3)We complete the word recognition of a group of keystrokes by using the word matching method,92% of the word recognition accuracy can be realized by using the relevance of the keystrokes in the word input.