查看文章

In this book chapter, grant‐free uplink transmission in an Internet of Things (IoT) network that comprises a base station (BS) with multiple antennas and massive single‐antenna IoT devices is studied. Given the intermittent data traffic, we approach the challenge by formulating the joint activity detection and channel estimation (JADCE) problem as a group–sparse matrix recovery problem. While existing compressed sensing methods can address this formulated problem, they often grapple with the constraint of high computational complexity. Therefore, by leveraging algorithm unrolling, we propose an unrolled recurrent neural network (RNN) based on iterative shrinkage thresholding algorithm to simultaneously reduce computational complexity and improve matrix recovery performance. Recognizing that a nonconvex sparsity‐induced penalty can more effectively encourage …