Gesture Classification Based On Channel State Information

Gesture Classification Based On Channel State Information

연구 분야: Infrastructure

학회: SPML '24: Proceedings of the 2024 7th International Conference on Signal Processing and Machine Learning

Wireless technology has rapidly increased in popularity in the past few years in the research field of human gesture recognition with its advantages of providing high throughput. Channel state information measured by wireless networks is used for different sensing purposes. The objective of this study is to develop an effective gesture recognition system utilizing a monitor to monitor based setup based featuring three receivers and transmitters. By meticulously selection the most relevant data, employing sophisticated signal techniques including hampel filter coupled with Gaussian filter, segmented the data via peak detection, extracted the features via mean magnitude across frequency bins for every time frame this method is derived from Fast Fourier transform(FFT), finally coupled with CNN, (support vector machine)SVM for classification therefore enhancing human-computer interaction in diverse applications. Experimental results demonstrate that the proposed methodology in gesture recognition utilizing the selected data and combination of the signal processing techniques feature extraction and CNN classification, our system achieved (94.1%,92.7%)(SVM) and (98.4%,98.28%)(CNN) in classifying human hand gestures for both training and testing respectively.