Advancing Wireless IoT Sensor Networks With Edge-Powered RFID Devices

In the era of the Internet of Things (IoT), the drive to interconnect trillions of devices has placed significant emphasis on the wireless domain. To address the challenge of balancing power efficiency, reliability, efficient data transmission, affordability, and implementation ease, we propose a passive Radio Frequency Identification (RFID) based IoT wireless sensor network- an interconnected battery-less and wireless sensing device platform. Our solution consists of an array of virtually controlled computational RFID (VCRFID) sensor tags interfaced and powered wirelessly by a reader. We demonstrate efficient sensor data collection and processing for system monitoring and the implementation of an LSTM autoencoder for anomaly detection. The anomaly detection system combines a wireless sensor network of VCRFID tags and machine learning at the edge, ensuring data remains local, minimizing delays, and enabling power-efficient, wireless, edge-powered processing.