Fast MAPE-Based Device Activity Detection for Massive Grant-Free Access via PSCA and PSCA-Net

Fast and accurate device activity detection is a critical challenge in grant-free access to support massive machine-type communications (mMTC). The state-of-the-art methods have unsatisfactory error rates or computation times. To address these outstanding issues, we propose new maximum a posterior estimation (MAPE)-based device activity detection methods that achieve superior error rate and computation time tradeoffs using optimization and deep learning techniques. Specifically, we propose an innovative parallel successive convex approximation (PSCA) algorithm to solve a novel non-convex MAPE problem for device activity detection. We show that PSCA allows parallel computations, uses up to the objective function’s second-order information, converges to the MAPE problem’s stationary points, and has a low per-iteration computational complexity. Then, we propose a PSCA-driven deep unrolling neural network implementation, called PSCA-Net, to reduce the overall computation time of PSCA. PSCA-Net elegantly marries PSCA’s parallel computation mechanism with the parallelizable neural network architecture and effectively optimizes its step sizes based on vast data samples. Numerical results demonstrate that PSCA and PSCA-Net can significantly reduce the error rate and computation time compared to the state-of-the-art methods, revealing their significant values for mMTC in 5G and beyond.