The Ultra-Wideband Communication System: A Human Gesture Recognition Approach
View/ Open
Pagination
825 - 830
ISBN-13
9781665494571
DOI
10.1109/HPCC-DSS-SmartCity-DependSys53884.2021.00133
Metadata
Show full item recordAbstract
In this paper, we propose a cost-effective ultra-wideband (UWB) communication system for gesture recognition in a smart home environment, where the interference issues can be beneficially solved. In the proposed UWB communication system, the gesture trajectories obtained by positioning are employed for recognizing human gestures. To this end, we firstly collect the datasets of four different fine-grained gesture activities. Then, we integrate the squeeze-and-excitation block into the convolutional neural network seamlessly for gesture recognition, namely the SE-ConvlD model, which achieves an overall accuracy of 99.48%.