Radio Frequency Identification Vibration Detection Using Long Short-Term Memory

Barrett Durtschi, Janita Aamir, Dr. Paul Bodily, and Dr. Andrew Chrysler, Department of Electrical Engineering and Computer Science, Idaho State University, 921 S. 8th Avenue, Pocatello ID 83209

 Radio frequency identification (RFID) is a low-cost and important part of the Internet of Things (IoT) technology and has applications in healthcare, inventory management, object detection, tracking, and more. Due to these many applications and attractive cost, RFID technology is being investigated for possible applications in smart-infrastructure such as vibration detection.  In this research, we present a method of detecting vibration frequency using RFID technology. By implementing a Low-Level Reader Protocol (LLRP), we can receive the phase angle along with the received signal strength indicator (RSSI) that can be used for further analysis. We have chosen to execute our analysis through the use of a LSTM (Long Short-Term Memory) Network. LSTM Networks, a type of Recurrent Neural Network, are good for both long-term and short-term time-series predictions. The use of LSTM Network, in this application, will enable us to impute vibration frequencies from phase angle and RSSI data.  This will help analyze the pattern of previously recorded RFID data readings and make reliable inferences for vibration frequencies. To make a situation in which the RFID data can be interpreted in an LSTM, we use a shake table with a piece of mounted concrete to simulate vibration. RFID data is then gathered by setting the movement of the shake table at several different frequencies. We then train an LSTM model on the recorded and labeled RFID data and use cross-validation to evaluate the network’s accuracy at imputing vibration frequency from RFID data.

Additional Abstract Information

Presenters: Barrett Durtschi, Janita Aamir

Institution: Idaho State University

Type: Poster

Subject: Electrical & Computer Engineering

Status: Approved

Time and Location

Session: Poster 6
Date/Time: Tue 2:00pm-3:00pm
Session Number: 4527