This paper considers joint active user detection (AUD) and channel estimation (CE) for massive connectivity scenarios with sporadic traffic. The state-of-art method under a Bayesian framework to perform joint AUD and CE in such scenarios is approximate message passing (AMP). However, the existing theoretical analysis of AMP-based joint AUD and CE can only be performed with a given fixed point of the AMP state evolution function, lacking the analysis of AMP phase transition and Bayes-optimality. In this paper, we propose a novel theoretical framework to analyze the performance of the joint AUD and CE problem by adopting the replica method in the Bayes-optimal condition. Specifically, our analysis is based on a general channel model, which reduces to particular channel models in multiple typical MIMO communication scenarios. Our theoretical framework allows ones to measure the optimality and phase transition of AMP-based joint AUD and CE as well as to predict the corresponding performance metrics under our model. To reify our proposed theoretical framework, we analyze two typical scenarios from the massive random access literature, i.e., the isotropic channel scenario and the spatially correlated channel scenario. Accordingly, our performance analysis produces some novel results for both the isotropic Raleigh channel and spatially correlated channel case.
翻译:本文考虑对零星交通的大规模连通情况进行联合积极用户检测(AUD)和频道估计(CE ) 。巴伊西亚框架下在这类情况下联合实施AUD和CE的最新方法大致是传递信息(AMP ) 。然而,对AMP基础的联合AUD和CE的现有理论分析只能用AMP州演化功能的某个固定点来进行,缺乏对AMP阶段过渡和Bayes-优化的分析。本文提出一个新的理论框架,分析AUD和CE问题的联合性能,在巴伊斯-最佳条件下采用复制方法。具体地说,我们的分析基于一个通用的频道模型,该模型在多种典型的MIMIM通信情景中减少为特定的频道模型。我们的理论框架允许测量AMP联合AUD和CE的优化性和阶段性过渡,并预测我们模型下的相应性能衡量尺度。为了更新我们提议的理论框架,我们从大规模随机访问文献中分析两种典型的情景,即从一些Bayes-optropic 频道情景和空间频道分析结果。