Research And Implementation Of Terminal Gesture Control Method Of Intelligent Control

In previous researches on human machine interaction, the parameters or templates of gestures are always learnt from training samples first and then cer-tain kind of matching is conduct.For these training-required methods, a small number of training samples always result in poor or user-independent perfor-mance. While a large quantity of training samples lead to time-consuming and laborious sample collection process.In this paper, two high-performance training-free approach for hand ges-ture recognition with accelerometer is proposed. For the basic approach, it utilizes principle component analysis for energy distribution analysis to bypass the gratitude influence, and extracts the physical characteristics of the gesture trajectory for classification. In this real-time gesture recognition system, first the acceleration data of a gesture is collected through Bluetooth interface be-tween WiiMote and PC; then the energy distribution shape is calculated, then several physical trajectory characteristic is organized as motion feature cascade classification. The proposed method is compared with the published method under the Nokia dataset, whose recognition accuracy could achieve91.35%for user-independent situation. For the improve approach, we determine the under-lining space for gesture generation with the physical meaning of acceleration direction. Then, the template of each gesture in underlining space can be gen-erated from gesture trails, which are frequently provided in the instructions of gesture recognition devices. Thus, during gesture template generation process, the algorithm does not require training samples any more and fulfills training-free gesture recognition. After that, a feature extraction method, which trans-forms the original acceleration sequence into a sequence of more user-invariant features in underlining space, and a more robust template matching method, which is based on dynamic programming, are presented to finish the gesture recognition process and enhance the system performance. Our algorithm is tested in Nokia dataset whose recognition accuracy could achieve93.1%for user-independent situation, the training-free algorithm even shows better per-formance than tradition training-required algorithms of Hidden Markov Model (HMM) and Dynamic Time Warping (DTW).