Character Input Method Using Back Tap on Smartphone

Abstract

With the widespread adoption of smartphones, opportunities for character input using the smartphone are increasing. However, character input on a smartphone is challenging. For example, it is erroneous to input owing to the small size button and small size information screen owing to the large size character input screen. To solve these problems, input methods besides touch are being studied. For example, input using sight and gesture. The aim of this study is to develop a gesture input interface that uses a small motion, that is, back tap of the device input. Back tap indicates the motion of tapping the left and right sides of the back of the device once or twice. Furthermore, combinations that can be used for discrimination are considered. The three-axis sensor of acceleration and gyro on a smartphone is used to determine the movement. In addition, the sampling frequency is set to 200 Hz. Using Z-axis of acceleration and Y-axis of gyro, the sensor could determine the analysis acceleration and gyro data of back taps. In the experiment based on the analysis results, four patterns were discriminated. Accordingly, it was possible to determine all patterns with at least 90% accuracy. Because any pattern could be discriminated with high accuracy, we consider character input in consideration of affordances. In addition, the character input speed and accuracy are calculated.