Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Scalability and Implementation Aspects of Cell-Free Massive MIMO for ISAC

Stefano Buzzi, Carmen D'Andrea, Sergi Liesegang

TL;DR

This work addresses scalability of cell-free massive MIMO for integrated sensing and communications in large-area deployments with multiple targets. It introduces a target-centric extension to the user-centric CF-mMIMO architecture and a bi-static sensing paradigm to manage sensing regions without full duplex operation. A joint downlink signal model is derived and a GLRT-based per-region detector is proposed to enable scalable sensing while preserving communication performance. Simulation results show that the proposed UTC approach can achieve or closely approach the performance of non-scalable schemes with lower complexity, validating its practical potential.

Abstract

This paper addresses the problem of scalability for a cell-free massive MIMO (CF-mMIMO) system that performs integrated sensing and communications (ISAC). Specifically, the case where a large number of access points (APs) are deployed to perform simultaneous communication with mobile users and monitoring of the surrounding environment in the same time-frequency slot is considered, and a target-centric approach on top of the user-centric architecture used for communication services is introduced. In the paper, other practical aspects such as the fronthaul load and scanning protocol are also considered. The proposed scalable ISAC-enabled CF-mMIMO network has lower levels of system complexity, permits managing the scenario in which multiple targets are to be tracked/sensed by the APs, and achieves performance levels superior or, in some cases, close to those of the non-scalable solutions.

Scalability and Implementation Aspects of Cell-Free Massive MIMO for ISAC

TL;DR

This work addresses scalability of cell-free massive MIMO for integrated sensing and communications in large-area deployments with multiple targets. It introduces a target-centric extension to the user-centric CF-mMIMO architecture and a bi-static sensing paradigm to manage sensing regions without full duplex operation. A joint downlink signal model is derived and a GLRT-based per-region detector is proposed to enable scalable sensing while preserving communication performance. Simulation results show that the proposed UTC approach can achieve or closely approach the performance of non-scalable schemes with lower complexity, validating its practical potential.

Abstract

This paper addresses the problem of scalability for a cell-free massive MIMO (CF-mMIMO) system that performs integrated sensing and communications (ISAC). Specifically, the case where a large number of access points (APs) are deployed to perform simultaneous communication with mobile users and monitoring of the surrounding environment in the same time-frequency slot is considered, and a target-centric approach on top of the user-centric architecture used for communication services is introduced. In the paper, other practical aspects such as the fronthaul load and scanning protocol are also considered. The proposed scalable ISAC-enabled CF-mMIMO network has lower levels of system complexity, permits managing the scenario in which multiple targets are to be tracked/sensed by the APs, and achieves performance levels superior or, in some cases, close to those of the non-scalable solutions.
Paper Structure (8 sections, 16 equations, 6 figures)

This paper contains 8 sections, 16 equations, 6 figures.

Table of Contents

  1. Introduction
  2. System model and scalability aspects for ISAC
  3. Signal model and transceiver processing
  4. Signal model
  5. Sensing design
  6. Other implementation aspects
  7. Numerical Results
  8. Conclusions

Figures (6)

  • Figure 1: A pictorial representation of the considered scenario. Several APs cooperate to perform communication and sensing tasks. The sets ${\cal M}_{\mathbf{p}_l}\cap {\cal M}_{\rm tx}$ and ${\cal M}_{\mathbf{p}_l}\cap {\cal M}_{\rm rx}$ refer to the transmit and receive APs inspecting position $\mathbf{p}_l$ in the $l$-th sensing area, respectively. In the figure, the set ${\cal K}_m$ of UEs is served by the $m$-th AP and the set ${\cal M}_k$ of APs is serving the $k$-th UE.
  • Figure 2: Illustrative example of a deployment with $L = 4$ regions.
  • Figure 3: Illustrative example of a deployment with no sensing regions.
  • Figure 4: CDF of the UE data rate (left) and of the sensing SNR (right) under: (a) UTC, (b) UC, (c) TC, and (d) pure CF implementations.
  • Figure 5: CDF of the UE data rate (left) and of sensing SNR (right) for several numbers of receiving APs $M_{\rm rx} = \left|{\cal M}_{\mathbf{p}_l} \cap {\cal M}_{\rm rx}\right|$ with UTC.
  • ...and 1 more figures