Enabling Low-Power Massive MIMO with Ternary ADCs for AIoT Sensing DOI: 10.1145/3722220 Journal: TOSN Volume: 21 5 3 5 CCS: Networks Mobile networks CCS: Hardware Signal processing systems CCS: Hardware Wireless devices
Shengheng Liu, Ningning Fu · 2025
Read Enabling Low-Power Massive MIMO with Ternary ADCs for AIoT Sensing DOI: 10.1145/3722220 Journal: TOSN Volume: 21 5 3 5 CCS: Networks Mobile networks CCS: Hardware Signal processing systems CCS: Hardware Wireless devices by Shengheng Liu, Ningning Fu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-11234.txt.
Introduction
Massive antenna arrays can serve many devices at once, but their converters, which turn radio waves into numbers, eat power. Shengheng Liu and Ningning Fu explore in 2025 a leaner design that uses ternary converters, which report only three levels instead of many. Their target is the artificial intelligence of things, where countless small sensors must sip energy.
Cutting converter precision is a direct way to cut the power drawn by each radio chain. The authors weigh what that saving costs in sensing quality and how the trade can be managed. Their study speaks to networks that must run for years on tight energy budgets. It frames coarse conversion as a design choice rather than a flaw.
Catalog record
| CREATOR | Shengheng Liu, Ningning Fu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 99 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-11234
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Enabling Low-Power Massive MIMO with Ternary ADCs for AIoT Sensing DOI: 10.1145/3722220 Journal: TOSN Volume: 21 5 3 5 CCS: Networks Mobile networks CCS: Hardware Signal processing systems CCS: Hardware Wireless devices Shengheng Liu, Ningning Fu Abstract . The proliferation of networked devices and the surging demand for ubiquitous intelligence have given rise to the artificial intelligence of things (AIoT). However, the utilization of high-resolution analog-to-digital converters (ADCs) and numerous radio frequency chains significantly raises power consumption. This paper explores a cost-effective solution using ternary ADCs (T-ADCs) in massive multiple-input-multiple-output (MIMO) systems for low-power AIoT and specifically addresses channel sensing challenges. The channel is first estimated through a pilot-aided scheme and refined using a joint-pilot-and-data (JPD) approach. To assess the performance limits of this two-threshold ADC system, the analysis includes its hardware-ideal counterpart, the parallel one-bit ADCs (PO-ADCs) and a realistic scenario where noise variance is unknown at the receiver is considered. Analytical findings indicate that the JPD scheme effectively mitigates performance degradation in channel estimation due to coarse quantization effects under mild conditions, without necessitating additional pilot overhead. For deterministic and random channels, we propose modified expectation maximization (EM) and variational inference EM estimators, respectively. Extensive simulations validate the theoretical results and demonstrate the effectiveness of the proposed estimators in terms of mean square error and symbol error rate, which showcases the feasibility of implementing T-ADCs and the associated JPD scheme for greener AIoT smart sensing. 1. Introduction Mobile service providers rank among the largest energy consumers worldwide, responsible for up to three percent of the total global power usage. The growing increase in the contact-free sensing demands and densification of artificial intelligence of things (AIoT) facilities create substantial pressure for the service providers to adopt environmentally sustainable practices. This push is not only driven by ethical considerations but also by commercial imperatives. Given that the radio access network (RAN) accounts for the majority, 87\% as per ( Association 2023 ) , of telecom energy consumption, targeting RANs for green improvements is a logical priority ( Cao et al. 2019 ) . Given that a significant proportion of AIoT devices are low-power battery-driven devices, considerable research interest has been directed towards the base station (BS) side. Massive multiple-input multiple-output (MIMO) technology ( You et al. 2023 ; Yan et al. 2022 ; Nguyen and Krunz 2014 ) is capable of creating a large number of spatial degrees of freedoms, which are essential in achieving high spectral efficiency, high data rates, and spatial multiplexing. In typical massive MIMO setups, hundreds of antennas are employed, each connected to a dedicated radio frequency (RF) chain equipped with high-resolution (e.g., 12–-16 bits) analog-to-digital converters (ADCs) ( Chung and Ye 2019 ) . However, ADCs are known to consume a substantial amount of power dissipated at a base station receiver, and the power consumption grows exponentially with resolution level and linearly with sampling rate ( Rangan et al. 2014 ) . Consequently, high-resolution quantization poses a significant challenge to the energy efficiency of massive MIMO based AIoT systems is inviable commercially. While low-resolution ADC deployments enable economically attractive receivers, coarse quantization inevitably leads to a fundamental shift, resulting in highly distorted and nonlinear quantized signals. Consequently, using conventional linear estimators/detectors to handle quantized signals undergoes inherent performance degeneration ( Choi et al. 2016 ; Saxena et al. 2017 ) , especially in one-bit quantization systems. Channel estimators for quantized MIMO systems are mainly developed under the principle of maximum likelihood (ML) ( Choi et al. 2016 ; Mezghani and Swindlehurst 2018 ) or Bayes rules ( Wen et al. 2016 ; Ding et al. 2018 ; Zhang et al. 2017 ) in the literature. In ( Li et al. 2017 ) , a Bussgang decomposition based channel estimator for one-bit massive MIMO systems was developed. In ( Mezghani et al. 2010 ) , the ML estimates of multiple parameters of interest within quantized observations were obtained by the expectation maximization (EM) algorithm. In ( Wen et al. 2016 ) , the generalized approximate message passing based channel estimator was proposed for the quantized massive MIMO system, which has been proven to attain the optimal Bayes limit asymptotically, and its enhanced variant, namely vector approximate message passing algorithm, to deal with the broadband millimeter channel, was investigated in ( Mo et al. 2017 ) . In ( Ding et al. 2018 ) , a hybrid analog-digital preprocessing structure was integrated with the quantized mmWave massive MIMO system, which develops a super-resolution beamspace channel estimator under the sparse Bayesian learning framework. A more advanced antithetic dithered one-bit receiver was analyzed in ( Ho and Rao 2019 ) , and the proposed parameter expanded EM channel estimator has been verified to achieve fast convergence and monotonic MSE performance over a wide signal-to-noise ratio (SNR) range. Throughput analyses in works such as ( Singh et al. 2009 ; Mo and Jr. 2015 ; Jacobsson et al. 2017 ) and error bound studies in ( Li et al. 2017 ; Rao et al. 2019 ; Liu et al. 2020 ) reveal a nonnegligible and systematic performance loss in one-bit systems compared to their full-resolution counterparts. To mitigate this performance degradation without incurring substantial hardware overhead, the ternary ADC (T-ADC) architecture was developed in ( Tanoue et al. 2007 ; Zhang et al. 2020 ) . Unlike conventional integer-bit ADC, the T-ADC compares the analog input against two nonidentical thresholds, generating quantized outputs in three possible intervals. In massive MIMO systems, T-ADCs offer a middle ground in power consumption, sitting between one-bit and two-bit ADCs, yet surpass both in energy efficiency ( Tanoue et al. 2007 ; Zhang et al. 2020 ) . From a physical perspective, as a type of two-threshold ADC, the hardware-ideal counterpart of the T-ADC is the parallel one-bit ADC (PO-ADC), as depicted in Fig. 1 . This paper explores the challenge of channel estimation using low-resolution ADCs in AIoT sensing applications and provides theoretical error bounds concerning various system parameters. Additionally, we propose two channel estimators, one tailored for deterministic channels and the other for random channels, to boost transmission performance in massive MIMO systems for AIoT sensing. The technical contributions of this paper are summarized as follows. • We consider the situation where the noise variance is unknown at the receiver ( Liu et al. 2020 ; Ribeiro and Giannakis 2006 ) . For pilot-aided (PA) channel estimation, we derive the Cramér-Rao lower bounds (CRLBs) on the channel parameter estimates to determine the performance limit of such a quantized massive MIMO system. We also deduce analytically expressible CRLBs for the single-input-single-output (SISO) system, which explicitly describe how the relative distance between the ADCs’ thresholds and the channel magnitude affects the CRLBs. Analytical and numerical results show that the T-ADC achieves comparable error performance to the PO-ADC over a wide SNR range, with reduced power consumption. • We propose a joint-pilot-and-data (JPD) scheme to leverage all the transmitted symbols in channel estimation, significantly reducing pilot overhead, which can be prohibitive for massive MIMO systems. Due to the coarse quantization of the T-ADC, the robustness of the PA estimation heavily relies on the pilot length. Rigorous CRLB analysis shows that the JPD scheme is comparable to