Research On Accelerometer-based Gesture Recognition For The Android Platform

With the rapid development of microelectronics technology, intelli-gent terminals become smarter and they are playing a more and more important role in human’s daily life. The interactive mode between hu-man and intelligent terminals has become one of hottest fields of hu-man-computer interaction. It’s gradually emerging from a comput-er-centric interactive mode towards people-centered interactive mode. Gesture, as a nature and intuitive communication way, is widely applied in human-computer interaction.Android platform is a widely used open-source operating system for smartphone. It provides a great convenience to the HCI research on in-telligent terminals. Based on the research of related theories, we de-signed and implemented an accelerometer-based gesture recognition system on intelligent terminal with Android. In this system, the start and end of the data collection process is automatically determined by accel-eration waveform. In pretreatment phase, we propose a waveform com-pensation algorithm to solve the problems caused by the amplitude range of the accelerometer and use the coordinate transformation theory to alleviate the angle offset. In training phase, we use dynamic time warping (DTW) and affinity propagation (AP) to extract clusters and exemplars. We implement sparse representation for gesture recognition and propose a modified variable sparsity adaptive matching pursuit (MVSAMP) algorithm for signal reconstruction. This algorithm is more adapted to the characteristics of gesture recognition. In the classification stage, a method of weighted residuals is applied to improve the resolu-tion of the best classification.To test the system’s performance, a dictionary of10gestures is de-fined and a database consists of3800samples is created from14partici- pants. In user-dependent condition, the system has achieved an average recognition accuracy of99.67%. And it has achieved an average recogni-tion accuracy of97%in user-independent experiment. Test results have shown that the proposed system achieves a good performance in a variety of experiments on Android platform.