Power Allocation in HARQ-Based Predictor Antenna Systems

Hao Guo, Behrooz Makki, Mohamed-Slim Alouini, Tommy Svensson

TL;DR

This paper introduces a novel Hybrid Automatic Repeat Request (HARQ) protocol for Predictor Antenna (PA) systems on moving vehicles to combat spatial mismatch and improve spectral efficiency. By utilizing the front predictor antenna for both initial data transmission and channel estimation, and applying optimized power allocation for retransmissions, the proposed system reduces outage-limited power consumption by up to 30 dB compared to traditional setups without retransmission.

What Problem Does This Paper Address?

In 5G and beyond, providing reliable, high-speed connections to moving vehicles (cars, buses, trains) is a major priority. To do this efficiently, the Base Station (BS) needs accurate Channel State Information at the Transmitter (CSIT). However, as a vehicle speeds up, the channel changes rapidly, making standard CSIT outdated by the time data is transmitted.

To solve this, researchers use Predictor Antenna (PA) systems: a setup with two antennas aligned front-to-back on a vehicle's roof. The front antenna (the PA) measures the channel and passes that info to the BS, which then transmits data to the rear Receive Antenna (RA) when it reaches the exact same spatial position.

However, this system faces two major challenges:

  1. Spatial Mismatch: Due to varying vehicle speeds and BS processing delays, the RA might not perfectly align with the spatial point the PA sampled, leading to inaccurate CSIT and dropped packets.
  2. Spectral Underutilization: In standard setups, the front PA is only used for channel estimation, wasting a perfectly good antenna that could be receiving data.

Methods

The authors propose a HARQ-based PA system that puts the front antenna to work.

  1. Initial Transmission: The BS sends pilots and encoded data to the front PA. The PA estimates the channel and simultaneously tries to decode the data.
  2. Feedback: If the PA successfully decodes the data, it sends an ACK, and transmission stops. If it fails, it sends a NACK along with high-accuracy quantized CSI back to the BS.
  3. Retransmission with Power Allocation: The BS then retransmits the data to the rear RA. Crucially, the BS uses the CSI provided by the PA to dynamically adapt its transmission power to overcome any spatial mismatch.

The authors cast this as an optimization problem: minimizing the average transmission power subject to a strict outage probability constraint. Because the exact mathematical formulations involve complex integrals of the Marcum Q-function, the authors develop tight closed-form approximations using the Lambert W function. They evaluate this for two standard HARQ protocols: Repetition Time Diversity (RTD) and Incremental Redundancy (INR).

Datasets

Not discussed in the paper. The authors evaluate their proposed system using theoretical mathematical analysis and Monte Carlo simulations (averaging over $1 \times 10^5$ Rayleigh fading channel realizations) rather than empirical datasets.

Key Insights and Ideas

  • Dual-Purpose Antennas Boost Efficiency: Using the predictor antenna for both channel estimation and initial data transmission significantly boosts the spectral efficiency of the system without requiring extra hardware.
  • HARQ and PA are a Perfect Match: The physical time delay between the front PA and rear RA passing the same point in space perfectly accommodates the time needed for HARQ feedback and BS processing.
  • Power Allocation is the Real Hero: Simply adding HARQ to a PA system provides a moderate performance bump. However, dynamically adapting the retransmission power based on the predicted CSI yields massive, exponential gains, especially when spatial mismatch (due to high vehicle speeds) is severe.
  • INR Outperforms RTD: As is typical in HARQ systems, sending new redundancy bits (INR) outperforms simply repeating the same signal (RTD) in terms of outage-limited average power, though the performance gap narrows at lower data rates or lower spatial correlations.

Key Results

  • Massive Power Savings: At a strict outage probability constraint of $10^{-5}$ and an initial rate of 4 nats-per-channel-use (npcu), the power-adaptive PA-HARQ scheme reduces the required Signal-to-Noise Ratio (SNR) by 25 dB (for RTD) and 30 dB (for INR) compared to baseline systems with no retransmission.
  • Highly Accurate Approximations: The derived closed-form mathematical approximations for the minimum required power tightly match the numerical simulation results across a broad range of vehicle speeds, initial data rates, and outage constraints.

Limitations and Future Work

While the theoretical results are highly promising, the paper relies on a few simplifying assumptions. First, it assumes the BS receives perfect CSIT from the PA's feedback, assuming the quantization bits are large enough to ignore feedback errors. Second, the optimization problem ignores peak power constraints, assuming the BS is capable of transmitting at whatever high power the algorithm dictates to meet the outage constraint. Finally, the model assumes a semi-static propagation environment where the coherence time is much larger than the processing delay.

Future work could explore the impact of imperfect/limited CSI feedback, introduce strict peak power constraints, and test the protocol in highly dynamic, non-stationary fading environments.

Summary updated

Abstract

In this letter, we study the performance of predictor antenna (PA) systems using hybrid automatic repeat request (HARQ). Here, the PA system is referred to as a system with two sets of antennas on the roof of a vehicle. In this setup, the PA positioned in the front of the vehicle can be used to predict the channel state information at the transmitter (CSIT) for the receive antenna (RA) that is aligned behind the PA. Considering spatial mismatch due to mobility, we derive closed-form expressions for the optimal power allocation and the minimum average power of PA systems under different outage probability constraints. The results are presented for different types of HARQ protocols and we study the effect of different parameters on the performance of PA systems. As we show, our proposed approximation scheme enables us to analyze PA systems with high accuracy. Moreover, for different vehicle speeds, we show that HARQ-based feedback can reduce the outage-limited power consumption of PA systems by orders of magnitude.
BibTeX
@article{guo2020power,
  author = {Hao Guo and Behrooz Makki and Mohamed-Slim Alouini and Tommy Svensson},
  title = {{Power Allocation in HARQ-Based Predictor Antenna Systems}},
  journal = {IEEE Wireless Communications Letters},
  year = {2020},
  url = {https://openreview.net/forum?id=3Wlu5XMTJCb},
}

Loading related content...

Figure