共享码本随机接入有限块长信息理论极限分析

摘要：超大规模机器通信（umMTC）是6G移动通信网络的典型应用场景。针对umMTC场景中的共享码本随机接入问题展开信息理论研究，在给定基站天线数、码长、用户载荷和发送功率的情况下，推导了平均每用户误差概率的非渐近可达界和逆定理。仿真结果证实了可达界和逆定理的紧致性，并且证明了当用户数低于某一门限时存在多用户干扰消除效应。基站配置多天线能够显著降低平均每用户误差概率，这对支持大量用户可靠接入具有重要意义。此外，与独立码本随机接入相比，共享码本随机接入在能效性方面更具优势。

关键词：共享码本；大规模随机接入；有限块长；可达界和逆定理

Abstract: Ultra-massive Machine Type Communication (umMTC) is expected to be a key use case in the 6th Generation wireless networks.  The information-theoretical analysis for unsourced random access in umMTC scenarios is conducted. Thenon-asymptotic achievability and converse bounds are established on the per-user probability of error for given number of receive antennas, blocklength, payload size, and transmitting power. Simulation results verify the tightness of our bounds and demonstrate that when the number of users is below a threshold, the multi-user interference can be almost perfectly canceled. The per-user probability of error decreases as the number of receive antennas increases, highlighting the crucial role of MIMO in supporting reliable access for a large number of users. Moreover, it is shown that in contrast to random access with individual codebooks, random access with a common codebook achieves higher energy efficiency.

Keywords: common codebook; massive random access; finite blocklength; achievability and converse bounds