Sensors, Vol. 19, Pages 4086: Noninvasive Suspicious Liquid Detection Using Wireless Signals
Sensors, Vol. 19, Pages 4086: Noninvasive Suspicious Liquid Detection Using Wireless Signals
Sensors doi: 10.3390/s19194086
Authors:
Jiewen Deng
Wanrong Sun
Lei Guan
Nan Zhao
Muhammad Bilal Khan
Aifeng Ren
Jianxun Zhao
Xiaodong Yang
Qammer H. Abbasi
Conventional liquid detection instruments are very expensive and not conducive to large-scale deployment. In this work, we propose a method for detecting and identifying suspicious liquids based on the dielectric constant by utilizing the radio signals at a 5G frequency band. There are three major experiments: first, we use wireless channel information (WCI) to distinguish between suspicious and nonsuspicious liquids; then we identify the type of suspicious liquids; and finally, we distinguish the different concentrations of alcohol. The K-Nearest Neighbor (KNN) algorithm is used to classify the amplitude information extracted from the WCI matrix to detect and identify liquids, which is suitable for multimodal problems and easy to implement without training. The experimental result analysis showed that our method could detect more than 98% of the suspicious liquids, identify more than 97% of the suspicious liquid types, and distinguish up to 94% of the different concentrations of alcohol.
Source: Sensors - Category: Biotechnology Authors: Jiewen Deng Wanrong Sun Lei Guan Nan Zhao Muhammad Bilal Khan Aifeng Ren Jianxun Zhao Xiaodong Yang Qammer H. Abbasi Tags: Article Source Type: research