Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Sensing Management for Pilot-Free Predictive Beamforming in Cell-Free Massive MIMO Systems

Eren Berk Kama, Murat Babek Salman, Isaac Skog, Emil Björnson

TL;DR

This paper introduces a sensing management method for integrated sensing and communications (ISAC) in cell-free massive multiple-input multiple-output (MIMO) systems that provides uniform downlink communication rates that are higher than with existing methods by achieving overhead-free predictive beamforming.

Abstract

This paper introduces a sensing management method for integrated sensing and communications (ISAC) in cell-free massive multiple-input multiple-output (MIMO) systems. Conventional communication systems employ channel estimation procedures that impose significant overhead during data transmission, consuming resources that could otherwise be utilized for data. To address this challenge, we propose a state-based approach that leverages sensing capabilities to track the user when there is no communication request. Upon receiving a communication request, predictive beamforming is employed based on the tracked user position, thereby reducing the need for channel estimation. Our framework incorporates an extended Kalman filter (EKF) based tracking algorithm with adaptive sensing management to perform sensing operations only when necessary to maintain high tracking accuracy. The simulation results demonstrate that our proposed sensing management approach provides uniform downlink communication rates that are higher than with existing methods by achieving overhead-free predictive beamforming.

Sensing Management for Pilot-Free Predictive Beamforming in Cell-Free Massive MIMO Systems

TL;DR

This paper introduces a sensing management method for integrated sensing and communications (ISAC) in cell-free massive multiple-input multiple-output (MIMO) systems that provides uniform downlink communication rates that are higher than with existing methods by achieving overhead-free predictive beamforming.

Abstract

This paper introduces a sensing management method for integrated sensing and communications (ISAC) in cell-free massive multiple-input multiple-output (MIMO) systems. Conventional communication systems employ channel estimation procedures that impose significant overhead during data transmission, consuming resources that could otherwise be utilized for data. To address this challenge, we propose a state-based approach that leverages sensing capabilities to track the user when there is no communication request. Upon receiving a communication request, predictive beamforming is employed based on the tracked user position, thereby reducing the need for channel estimation. Our framework incorporates an extended Kalman filter (EKF) based tracking algorithm with adaptive sensing management to perform sensing operations only when necessary to maintain high tracking accuracy. The simulation results demonstrate that our proposed sensing management approach provides uniform downlink communication rates that are higher than with existing methods by achieving overhead-free predictive beamforming.

Paper Structure

This paper contains 14 sections, 27 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Communication and Sensing System Model
  3. Channel Model
  4. Downlink Data Transmission
  5. Sensing Signal Transmission
  6. Multi-Static Sensing and its Performance
  7. User Tracking
  8. Extended Kalman Filter
  9. Sensing Management
  10. Simulation Setting and Results
  11. Instantaneous Downlink Rates
  12. Simulation Results
  13. Conclusion
  14. Appendix

Figures (5)

  • Figure 1: Conceptual diagram of the proposed predictive beamforming method. Sensing and tracking provide an estimate of the user position information when needed and the precoder is formed on demand.
  • Figure 2: The considered system where the first AP is the transmitter (Tx) AP and a set of Rx APs are selected for sensing reception.
  • Figure 3: The conventional and proposed frame structures used in the transmission from the CPU to the UE. The sizes of blocks are chosen differently to imply the changing lengths of the states.
  • Figure 4: The temporal behavior of the predicted angle estimation variance and sensing decisions with optimal and random Rx AP selection.
  • Figure 5: The temporal behavior of the instantaneous rate with and without sensing management and the perfect angle estimates with $L=4$, $N=4$.