A Framework for Hardware Impairments-Aware Multi-Antenna Transceiver Design in IoT Systems via Majorization-Minimization

• Published in: IEEE internet of things journal Vol. 10; no. 1; pp. 417 - 433
• Main Authors: Gong, Shiqi, Wang, Jintao, Zhao, Xin, Ma, Shaodan, Xing, Chengwen
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Antenna design; Antennas; Average total mean square error (MSE); Bit error rate; Codes; Data transmission; Errors; Hardware; hardware impairments; Internet of Things; Iterative algorithms; Iterative methods; low-complexity scheme; Lower bounds; majorization–minimization (MM)-based iterative algorithm; MIMO communication; Minimization; MSE floor effect; Optimization; Receivers; stochastic channel state information (CSI) errors; Transceivers
• ISSN: 2327-4662; 2327-4662
• DOI: 10.1109/JIOT.2022.3200828

In view of the nonideality of communication links in the Internet of Things (IoT) originating from transceiver hardware impairments, in this article, we introduce a general framework for hardware impairments-aware multiantenna transceiver design, which considers different availabilities of CSI at the transmitter (CSIT) and the receiver (CSIR). The well-known Kronecker model is applied to characterize stochastic channel state information (CSI) errors. For each case, we aim to minimize the (average) total mean square error (MSE) of all data streams subject to the practical per-antenna power constraints. To address the nonconvexity of the formulated problem, we propose an efficient majorization-minimization (MM)-based iterative algorithm to transform the original problem into a series of convex subproblems with semiclosed-form optimal solutions. For low-complexity implementation, we also develop an alternative scheme for directly finding a high-quality suboptimal solution by considering both worst case hardware impairments and worst case CSI errors. In particular, since an explicit expression of the average total MSE for the perfect CSIR and imperfect CSIT case is hard to derive, we instead optimize its effective upper and lower bounds. The prospective applications of our work in the two currently popular multiple-input-multiple-output (MIMO) IoT scenarios are then discussed. Furthermore, we fundamentally reveal the MSE floor effect caused by both hardware distortion and CSI imperfection in the high-SNR regime. Numerical results illustrate the excellent average total MSE and average bit error rate (BER) performance of our proposed algorithms over the adopted benchmark schemes.