Research on active user detection and channel estimation algorithms in massive access

Abstract

Massive Access is a supporting technology for scenarios like the Internet of Things, and is also one of the typical scenarios defined by 5G. Unlike traditional multi-user communication, the data of these nodes is generally sudden, so only a small proportion of nodes are active at a certain time. If the traditional request resource allocation protocol mode is adopted, it requires authorized access to achieve dynamic resource allocation, which will result in excessive protocol overhead. Therefore, Grant-Free transmission is required. It is impossible to recover the data transmitted by active users during Grant-Free transmission because the base station is unaware of who is transmitting the data. Therefore, the system must first do active user detection and channel estimation. A channel estimation technique based on Compressed Sensing is presented in this paper. the EM-BG-AMP algorithm is reproduced and applied. By comparing the accuracy of the EM-BG-AMP algorithm in channel estimation under physical backgrounds, this essay gives the range of optimal received signal-to-noise ratio:100dB and more. Also, this essay demonstrates that the performance of the EM-BG-AMP algorithm weakens as sparsity increases.