Impact of Channel Aging and Electromagnetic Interference on RIS-Assisted Cell-Free Massive MIMO Systems

TL;DR

This work analyzes RIS-assisted cell-free massive MIMO under channel aging and EMI, introducing a two-phase channel estimation scheme and large-scale fading-based pilot assignment to mitigate performance degradation. It derives closed-form downlink spectral efficiency expressions and optimizes the RIS coefficient matrices using projected gradient ascent, applying fractional power control for robustness. Results show the two-phase estimator and RIS optimization yield meaningful SE gains (up to ~10–15% under moderate EMI), with larger RIS deployments providing more resilience to aging, though high EMI can limit benefits. The study provides practical design guidelines for 6G deployments that combine RISs with distributed APs under mobility and EMI conditions.

Abstract

Cell-free massive multiple-input multiple-output (MIMO) and reconfigurable intelligent surfaces (RISs) are two potential sixth-generation (6G) technologies. However, channel aging due to user mobility and electromagnetic interference (EMI) impinging on RISs can negatively affect performance. Existing research on RIS-assisted cell-free massive MIMO systems often overlooks these issues. This work focuses on the impact and mitigation of channel aging and EMI on RIS-assisted cell-free massive MIMO systems over spatially correlated channels. To mitigate the degradation caused by these issues, we introduce a novel two-phase channel estimation scheme with large-scale fading coefficient-aided pilot assignment to enhance channel estimation accuracy compared to conventional minimum mean square error estimators. We then develop closed-form expressions for the downlink spectral efficiency (SE) performance and using these, optimize the sum downlink SE with respect to the RIS coefficient matrices. This optimization is accomplished by the projected gradient ascent (GA) algorithm. The results show that our proposed two-phase channel estimation scheme can achieve a nearly 10%-likely SE improvement compared to conventional channel estimation in environments affected by channel aging. A further 10%~15%-likely SE improvement is achieved using the proposed GA algorithm compared to random RIS phases, especially when the number of RISs increases.

Figures (7)

• Figure 1: RIS-assisted cell-free massive MIMO systems with user mobility.
• Figure 2: NMSE vs $p_p$ with $\varsigma=20~\text{dB}$, $M=10$, $N=2$, $K=10$, $J=4$, $L=16$.
• Figure 3: SE vs the time instant index with $\varsigma=20~\text{dB}$, $M=10$, $N=2$, $K=10$, $J=4$, $L=16$.
• Figure 4: Sum SE vs the number of APs with $\varsigma=20~\text{dB}$, $N=2$, $K=10$, $J=4$, $L=16$.
• Figure 5: Average SE vs the number of antennas per AP with $\varsigma=20~\text{dB}$, $M=10$, $J=4$, $L=16$.