TL;DR
This paper tackles the challenge of quantifying the actual performance gains of full-duplex (FD) wireless communication in practical scenarios, moving beyond idealized assumptions. By developing a realistic model for self-interference (SI) in compact FD receivers and analyzing both single-channel and multi-channel OFDM systems, the authors provide explicit bounds on rate gains and propose efficient resource allocation algorithms.
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 immense self-interference (SI) caused by the transmitter's signal leaking into the receiver. While significant progress has been made in canceling this SI, achieving perfect cancellation is extremely difficult, and the impact of imperfect SI on actual rate gains remains poorly understood. Existing research often assumes perfect SI cancellation or uses simplified models that don't capture the nuances of practical FD devices, especially compact ones like smartphones. This paper aims to bridge this gap by providing a rigorous analysis of FD rate gains under realistic SI conditions, considering both single-channel and multi-channel OFDM systems.
Methods
The authors introduce a novel, frequency-selective model for residual SI in compact FD receivers, motivated by the characteristics of RF integrated circuits. This model accounts for the fact that SI cancellation might not be uniform across all frequencies. They then analyze three FD use cases: single-channel bidirectional, two single-channel unidirectional, and multi-channel bidirectional OFDM.
For resource allocation (power and frequency selection), they develop algorithms that aim to maximize the sum of uplink (UL) and downlink (DL) rates. Key algorithmic contributions include:
- Single-Channel Analysis: Proving that for FD to offer a rate gain, it's optimal to transmit at maximum power. They also derive conditions for the sum rate function to be "biconcave," a property crucial for optimization.
- Multi-Channel OFDM: Developing a polynomial-time algorithm that converges to a near-optimal solution for power allocation and maximum SI frequency selection.
- High SINR Approximation: For faster computation, they derive an efficient, near-optimal algorithm under a high Signal-to-Interference-plus-Noise Ratio (SINR) approximation.
Datasets
The paper utilizes measurements from their own RFIC FD receiver design ([25], [26]) to validate their frequency-selective SI model. No new benchmarks are introduced; instead, the focus is on analytical modeling and algorithmic development.
Key Insights and Ideas
- Realistic SI Matters: The frequency-selective nature of SI cancellation in compact FD receivers significantly impacts achievable rate gains, especially in OFDM systems.
- Power Allocation Simplicity: In scenarios where FD offers non-negligible gains, simple power allocation strategies (like equal power distribution or the high SINR approximation) are often near-optimal.
- Biconcavity is Key: Establishing conditions for the sum rate function to be biconcave is crucial for developing efficient optimization algorithms. This property is linked to the quality of SI cancellation relative to wireless channel gains.
- FD Extends Capacity Region: FD communication demonstrably expands the achievable rate region compared to traditional Time Division Duplex (TDD) systems, particularly when signal-to-noise ratios are sufficiently high.
Key Results
- Explicit Rate Gain Bounds: The paper provides analytical guarantees on FD rate gains as a function of receiver SNRs and SI profiles.
- Condition for Gains: A sufficient condition is derived for achieving positive rate gains from FD, relating SI cancellation levels to wireless channel gains.
- Algorithm Performance: The proposed algorithms are shown to converge to near-optimal solutions and are efficient, especially the high SINR approximation.
- Numerical Validation: Simulations and comparisons with measured data demonstrate significant capacity region extensions and rate improvements due to FD, especially at higher SNRs. For instance, at high SNRs, FD can utilize almost all channels in full-duplex mode.
Limitations and Future Work
- Compact Receiver Focus: The primary SI model is tailored for compact FD receivers. While the BS model is simpler, the detailed analysis focuses on the MS.
- MIMO Extension: The current work focuses on single-antenna systems. Generalizing the results to MIMO FD systems is a significant area for future research.
- Dynamic Environments: The analysis assumes static channel conditions. Extending the framework to dynamic environments with varying channel states and interference levels would be valuable.
- Scheduling and MAC Layer: The paper lays groundwork for FD MAC layer algorithm design, but further work is needed on scheduling algorithms that can handle both half-duplex and full-duplex users efficiently.
Summary updated