Space-Time Coded Over-the-Air Computation With Receive Diversity for 6G Massive IoT Networks

This paper proposes a novel space-time line-coded over-the-air computation technique (STLC-AirComp) for massive Internet-of-Things (IoT) sensing networks. The STLC-AirComp technique enables efficient computation of objective functions by exploiting a simple linear channel inversion structure between IoT sensors and an access point (AP) in single-input multiple-output (SIMO) networks. In particular, this technique enhances the performance of function computation, including metrics such as mean squared error (MSE), without the need for global channel state information (CSI) at the access point (AP). Additionally, STLC-AirComp allows function computation to be performed in a one-shot manner, eliminating the need for complex learning or optimization processes. We perform a rigorous analysis to derive the exact MSE performance of STLC-AirComp, providing fundamental insights into computational reliability for massive IoT networks. Within the STLC-AirComp framework, we show that the MSE approaches zero as the SNR increases sufficiently. Finally, Monte Carlo simulations validate our mathematical analysis, demonstrating strong agreement between the analytical and simulation results.