Resource Allocation and Rate Gains in Practical Full-Duplex Systems

Jelena Marasevic, Jin Zhou, Harish Krishnaswamy, Yuan Zhong, Gil Zussman

TL;DR

This paper provides a comprehensive analysis of the achievable rate gains from full-duplex (FD) wireless communication, considering practical limitations like imperfect self-interference (SI) cancellation. It introduces a new, measurement-validated model for SI in compact FD receivers and develops algorithms to optimize resource allocation for maximizing throughput in both single- and multi-channel FD systems.

What Problem Does This Paper Address?

Full-duplex (FD) communication, which allows simultaneous transmission and reception on the same frequency, promises to significantly boost wireless network throughput. However, a major hurdle is the strong self-interference (SI) signal that leaks from the transmitter to the receiver, requiring substantial cancellation (often over 100dB). While FD has been demonstrated, its actual rate gains under realistic SI conditions, especially in compact devices like smartphones, remain poorly understood. Existing research often assumes perfect SI cancellation or uses simplified models, failing to capture the nuances of frequency-selective SI inherent in practical RF front-ends. This paper aims to bridge this gap by providing a rigorous analysis of FD rate gains, developing realistic models, and proposing efficient resource allocation strategies for practical FD systems.

Methods

The authors tackle the problem by first analyzing single-channel FD scenarios, deriving conditions for rate gains and establishing that maximizing transmission power is often optimal. They then extend this to multi-channel FD systems, introducing a novel, frequency-selective SI model for compact FD receivers, validated by measurements. This model accounts for the fact that SI cancellation effectiveness can vary across frequencies, a crucial aspect for real-world devices. Based on this model, they formulate a sum-rate maximization problem and develop two key algorithms:

  1. MAXIMUMRATE: A polynomial-time algorithm that finds a near-optimal solution for resource allocation (power and SI cancellation frequency) in the general SINR regime, guaranteed to converge to a stationary point that is practically a global maximum under mild conditions.
  2. HSINR-MAXIMUMRATE: An efficient algorithm for the high SINR regime, which significantly reduces computation time while providing near-optimal results, and also identifies that placing the peak SI cancellation frequency in the middle of the band is optimal.

Datasets

The paper utilizes measurement data from their own developed compact FD receiver to validate the proposed frequency-selective SI model. No new benchmark datasets are introduced.

Key Insights and Ideas

  • Power Maximization is Often Optimal: In single-channel FD, if FD offers a rate gain over TDD, it's often achieved by simply transmitting at maximum power. This is a surprisingly simple result given the complexity of the problem.
  • Frequency-Selective SI is Crucial: The paper highlights that the frequency-dependent nature of SI cancellation is a critical factor in practical FD systems, especially for compact devices. Their new model captures this effectively.
  • Biconcavity is Key: Establishing conditions for the sum-rate function to be biconcave (concave in two sets of variables) is crucial for developing efficient optimization algorithms. This property allows for guaranteed convergence to optimal or near-optimal solutions.
  • Simple Power Allocation Works Well: For multi-channel FD, especially in the high SINR regime, simple power allocation strategies (like equal power distribution or the high SINR approximation) perform nearly as well as complex, optimized schemes when FD provides significant rate gains.
  • FD Gains Depend on SNR and SIC Quality: Numerical results show that substantial FD rate gains are typically achieved only when the signal-to-noise ratio (SNR) is sufficiently high and the self-interference cancellation is effective.

Key Results

  • The paper provides analytical conditions for achieving positive rate gains in FD systems, even with imperfect SI cancellation.
  • A new, measurement-validated model for frequency-selective SI in compact FD receivers is introduced.
  • Two algorithms (MAXIMUMRATE and HSINR-MAXIMUMRATE) are developed for optimal power allocation and SI cancellation frequency selection in multi-channel FD systems, offering polynomial time complexity and near-optimal performance.
  • Numerical evaluations demonstrate significant capacity gains from FD, particularly in scenarios with high SNRs and effective SI cancellation. For instance, with sufficient SNR, FD can utilize nearly all channels in full-duplex mode.
  • The study shows that simple power allocation strategies are near-optimal for FD when rate gains are substantial.

Limitations and Future Work

  • MIMO Systems: The current work focuses on single-antenna FD systems. Generalizing the results to MIMO FD systems is identified as a significant area for future research.
  • Channel-Dependent SIC: The paper assumes the impact of SIC is uniform across channels for the MS in multi-channel scenarios. Future work could explore scenarios where SIC effectiveness varies significantly across channels.
  • Scheduling and Multiple Access: Developing scheduling algorithms that efficiently manage both half-duplex and full-duplex users in mixed networks is highlighted as an important future direction.
  • Practical Implementation Details: While the SI model is validated, further research into the practical implementation challenges of advanced SIC techniques and their impact on rate gains is suggested.

Summary updated

Abstract

Full-duplex (FD) communication has the potential to substantially increase the throughput in wireless networks. However, the benefits of FD are still not well understood. In this paper, we characterize the FD rate gains in both single-channel and multi-channel use cases. For the single-channel case, we quantify the rate gain as a function of the remaining self-interference (SI) and signal-to-noise ratio values. We also provide a sufficient condition under which the sum of uplink and downlink rates on an FD channel is biconcave in the transmission power levels. Building on these results, we consider the multi-channel case. For that case, we introduce a new realistic model of a compact (e.g., smartphone) FD receiver and demonstrate its accuracy via measurements. We study the problem of jointly allocating power levels to different channels and selecting the frequency of maximum SI suppression, where the objective is to maximize the sum of the rates over uplink and downlink orthogonal frequency division multiplexing channels. We develop a polynomial time algorithm, which is nearly optimal, in practice, under very mild restrictions. To reduce the running time, we develop an efficient nearly optimal algorithm under the high SINR approximation. Finally, we demonstrate via numerical evaluations the capacity gains in different use cases and obtain insights into the impact of the remaining SI and wireless channel states on the performance.
BibTeX
@inproceedings{marasevic2015resource,
  author = {Jelena Marasevic and Jin Zhou and Harish Krishnaswamy and Yuan Zhong and Gil Zussman},
  title = {{Resource Allocation and Rate Gains in Practical Full-Duplex Systems}},
  booktitle = {SIGMETRICS},
  year = {2015},
  doi = {10.1145/2745844.2745872},
}
Figure