Intelligent Beam Tracking in Radar-Assisted MIMO-OFDM Communication Systems

When the base station (BS) communicates with terminals via narrow beams, the BS should have beam tracking capability due to the uncertainty of the terminals' locations especially in the high mobility scenario. Unlike conventional beam tracking based on pilots, we investigate beam tracking in th...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on vehicular technology 2024-11, Vol.73 (11), p.16774-16789
Main Authors: Wang, Yijie, Lou, Mengting, Qian, Wanyun, Bai, Yechao, Tang, Lan, Liang, Ying-Chang
Format: Article
Language:English
Subjects:
Algorithms
Antenna arrays
Array signal processing
Beam tracking
Communications systems
Cramer-Rao bounds
Extended Kalman filter
extended kalman filter (EKF)
Lower bounds
Machine learning
Millimeter wave communication
MIMO communication
OFDM
Orthogonal Frequency Division Multiplexing
Performance enhancement
posterior Cramer-Rao lower bound (PCRLB)
Radar beams
Radar tracking
reinforcement learning
resource allocation
Resource management
Training
Transmission
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:When the base station (BS) communicates with terminals via narrow beams, the BS should have beam tracking capability due to the uncertainty of the terminals' locations especially in the high mobility scenario. Unlike conventional beam tracking based on pilots, we investigate beam tracking in the framework of integrated communication and sensing. In the proposed architecture, the transmitting BS and the receiving BSs cooperate to complete target location and beam tracking. When the transmitting BS transmits to terminals, multiple receiving BSs apply extended kalman filter (EKF) or unscented kalman filter (UKF) to predict locations of terminals based on received echoes, and then the location information is fed back to the transmitting BS to assist in beam tracking. To improve the performance of beam tracking, we first analyze the posterior Cramer-Rao lower bound (PCRLB) of prediction performance under the first-order Taylor approximation. Then, we investigate the subcarrier and power allocation problem during tracking, which is essentially a sequential decision problem. We propose a novel algorithm that combines optimization and reinforcement learning (RL) to achieve optimal overall performance. Simulation results demonstrate the tracking performance and transmission rate of the proposed method, as well as its advantages compared to existing schemes.
ISSN:0018-9545
1939-9359
DOI:10.1109/TVT.2024.3419000