Predictor Antenna Systems: Exploiting Channel State Information for Vehicle Communications
Licentiate thesis, 2020

Vehicle communication is one of the most important use cases in the fifth generation of wireless networks (5G).  The growing demand for quality of service (QoS) characterized by performance metrics, such as spectrum efficiency, peak data rate, and outage probability, is mainly limited by inaccurate prediction/estimation of channel state information (CSI) of the rapidly changing environment around moving vehicles. One way to increase the prediction horizon of CSI in order to improve the QoS is deploying predictor antennas (PAs).  A PA system consists of two sets of antennas typically mounted on the roof of a vehicle, where the PAs positioned at the front of the vehicle are used to predict the CSI observed by the receive antennas (RAs) that are aligned behind the PAs. In realistic PA systems, however, the actual benefit is affected by a variety of factors, including spatial mismatch, antenna utilization, temporal correlation of scattering environment, and CSI estimation error. This thesis investigates different resource allocation schemes for the PA systems under practical constraints, with main contributions summarized as follows.

First, in Paper A, we study the PA system in the presence of the so-called spatial mismatch problem, i.e., when the channel observed by the PA is not exactly the same as the one experienced by the RA. We derive closed-form expressions for the throughput-optimized rate adaptation, and evaluate the system performance in various temporally-correlated conditions for the scattering environment. Our results indicate that PA-assisted adaptive rate adaptation leads to a considerable performance improvement, compared to the cases with no rate adaptation. Then, to simplify e.g., various integral calculations as well as different operations such as parameter optimization, in Paper B, we propose a semi-linear approximation of the Marcum Q-function, and apply the proposed approximation to the evaluation of the PA system. We also perform deep analysis of the effect of various parameters such as antenna separation as well as CSI estimation error. As we show, our proposed approximation scheme enables us to analyze PA systems with high accuracy.

The second part of the thesis focuses on improving the spectral efficiency of the PA system by involving the PA into data transmission. In Paper C, we analyze the outage-limited performance of PA systems using hybrid automatic repeat request (HARQ). With our proposed approach, the PA is used not only for improving the CSI in the retransmissions to the RA, but also for data transmission in the initial round.  As we show in the analytical and the simulation results, the combination of PA and HARQ protocols makes it possible to improve the spectral efficiency and adapt transmission parameters to mitigate the effect of spatial mismatch.

rate adaptation


outage probability

Beyond 5G

spatial correlation

predictor antenna (PA)

wireless backhaul


temporal correlation

channel state information (CSI)

integrated access and backhaul (IAB)

hybrid automatic repeat request (HARQ)


mobile relay

Marcum Q-function


Opponent: Dr. Dinh-Thuy Phan Huy, Orange Labs, France


Hao Guo

Chalmers, Electrical Engineering, Communication and Antenna Systems, Communication Systems

Fifth Generation Communication Automotive Research and innovation (5GCAR)

European Commission (Horizon 2020), 2017-06-01 -- 2019-05-31.

Areas of Advance

Information and Communication Technology

Subject Categories


Communication Systems

Signal Processing


Chalmers University of Technology


Opponent: Dr. Dinh-Thuy Phan Huy, Orange Labs, France

More information

Latest update