Seamless Physical-Layer Cross-Technology Communication from ZigBee to LoRa via Neural Networks

LoRa, designed for Low-Power, Wide-Area Networks (LPWANs), is widely used in the Internet of Things (IoT). In contrast, Wireless Personal Area Network (WPAN) technologies like ZigBee struggle to connect directly to LPWANs due to their limited communication range and differing modulation schemes. ZigBee uses Offset Quadrature Phase-Shift Keying (OQPSK) modulation, while LoRa employs Chirp Spread Spectrum (CSS) modulation, complicating cross-technology communication. To address this challenge, we propose a novel approach for seamless physical-layer cross-technology communication between ZigBee and LoRa networks, bridging the gap between short-range and long-range communication technologies. We introduce ZigRa, a communication method that leverages neural networks for efficient modulation translation between ZigBee's IEEE 802.15.4 standard and LoRa's CSS modulation. The core of ZigRa is a deep learning model that adapts and optimizes the transformation of ZigBee signals into ultra-narrowband single-tone sinusoidal signals, which can be reliably detected by LoRaWAN base stations. Our solution enables ZigBee devices to seamlessly connect to LoRa-based LPWANs, overcoming modulation mismatches and providing long-range connectivity. Extensive evaluations with both USRP hardware and commercial devices demonstrate that ZigRa achieves a frame reception rate exceeding 85% at distances up to 500 meters, significantly enhancing the interoperability and coverage of heterogeneous IoT networks.