Designed And Implement Of The Gesture Recognition System Based On Inertial Sensor

With the constantly development of Intellectual Hardware and Mobile Device, Natural Human-Computer Interaction technology has become an significant direction in this research field. Image-based recognition and Voice-based recognition, with the feature of convenient and high efficiency, have become a mature way of human-computer interaction and achieved some results. Gestures, with its intuitive features, gradually become the focus of mobile device researchers in recent years. In the field of human-computer interaction represented by gesture recognition at present, there are three main technical routes: recognition technology based on visual, gesture recognition based on radar wave and recognition technology based on inertial sensors. The development of these three technologies go hand in hand, and have been widely used in the Internet of things such as intelligent household and health monitoring.Different from the other two kinds of gesture recognition methods, the research contents of this paper is based on the inertial sensors and focus on gesture recognition technology according to redefined gesture set. This research, mainly in order to improve the accuracy of gesture recognition and applied to the actual scene, can achieve efficient human-computer interaction and improve user experience. This paper has completed following works:(1) Described the system plan according to actual application scenarios. First of all, according to the data which predefined gesture sets gather from collection of sensor, extracts characteristic quantity that doesn't change with individual differences and reflect gesture peculiarity, and make it use as the basis of gesture judgement; Secondly, focus on the analyzing of existing algorithm, and obtain the results; At last, analyze and make comparison of these several typical sensor module and communication module, finally build hardware platform which is suitable for actual scene.(2) On the basis of completing the whole hardware platform, to enhance the accuracy of recognition algorithm, this article use median filtering and low pass filtering processing to the random noise and impulse noise separately, capture real-time gesture data and write gesture recognition algorithm based on this feature.(3) The actual application is the most important step to test whether the gesture recognition algorithm can be applied to the human-computer interaction field. This paper make the gesture recognition system applied in actual PPT page turning scenario, verify its accuracy through repeat test of large number of multiple users, and finally confirm the feasibility that algorithm can recognize predefined gesture.The recognition system focused on predefined gesture that based on inertial sensor, which is designed in this paper, provides a certain reference value to the research and practical use of gesture recognition area.