Robust and Energy-Efficient Channel Estimation in RF Backscatter Tag-to-Tag Network

RF tag networks that utilize fully passive tags and backscatter communication offer a promising solution for ultra-low-power and cost-effective Internet of Things (IoT) applications. Accurate estimation of tag-to-tag channel state information (CSI), particularly channel phase and path loss, is critical for enabling a wide range of such applications. However, the absence of active radio IQ demodulation and the limited computational resources in passive tags pose significant challenges for precise CSI estimation. To address these challenges, we propose a multi-phase modulator tag architecture and two optimized low-complexity channel estimation techniques: the Linear Least Squares (LLS) and the Three Reflection Loads (TRL) technique. Both techniques are designed to minimize estimation error while maintaining ultra-low computational cost, making them well-suited for energy-constrained implementations. We validate the proposed techniques through measurements conducted in both indoor and outdoor environments with a discrete-component, battery-powered RF tag prototypes. Experimental results show that the 90th percentile phase estimation error is 11° in outdoor settings and 25° in indoor settings under ambient RF excitation at 915 MHz. These phase errors correspond to ranging inaccuracies of approximately 10 mm and 23 mm, respectively, over tag-to-tag distances from 28 cm to 228 cm. These results highlight the robustness, energy efficiency, and scalability of the proposed channel estimation techniques for practical deployment in passive RF tag networks.