期刊
出版社
IEEE
DOI: 10.1109/APWCS50173.2021.9548761
关键词
Micro Doppler signatures; Software defined radio; Fall Detection; convolutional neural network (CNN)
Microwave-based radar sensors are increasingly used in healthcare and security applications. The software-defined implementation allows detection and classification of different types of motions, suitable for elderly care and monitoring, as well as seeing through visually opaque materials for security applications. Investigating the use of micro-Doppler signatures for localizing and detecting human micro-motions, a convolutional neural network is trained to identify fall events after augmenting the micro-Doppler data.
Microwave-based radar sensors are increasingly been used for healthcare and security applications. The software defined implementation of the radars allows fall detection and classification of different types of motions enabling elderly care and monitoring without privacy invading cameras. In addition, such radar sensor allow seeing through visually opaque materials suitable for security applications. This paper investigates the use of micro-Doppler signatures of slowly moving objects to localize and detect and classify human micro-motions. Using the NI SDRs, we measure micro-Doppler signatures of various human motion scenarios. Thereafter, the micro Doppler signatures' data is augmented before being used to train a convolutional neural network that detects and identifies the fall events.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据