Source-linked AI summary
Self-Interference Cancellation with Nonlinear Distortion Suppression for Full-Duplex Systems
Elsayed Ahmed, Ahmed M. Eltawil, Ashutosh Sabharwal
TL;DR
Strong self-interference makes transmitter and receiver nonlinearities a limit on cancellable interference in full-duplex systems. The paper jointly estimates the self-interference channel and nonlinearity coefficients for distortion suppression, achieving performance less than 0.5 dB off a linear full-duplex system after three to four iterations.
Problem
Strong self-interference makes nonlinear distortion a limit on cancellable self-interference power in full-duplex systems.
Source: §I · p. 1Method
The digital-domain scheme suppresses transmitter and receiver nonlinear distortion using iterative joint estimation of the self-interference channel and nonlinearity coefficients.
Source: §I · p. 1Results
After three to four iterations, performance is less than 0.5 dB off that of a linear full-duplex system.
Source: §I · p. 1Takeaways & Limitations
When nonlinear distortion dominates other noise components, distortion suppression allows full-duplex systems to achieve better rates than half-duplex systems at high SNR.
Source: §IV · pp. 4–5
Abstract
from arXiv · showhide
In full-duplex systems, due to the strong self-interference signal, system nonlinearities become a significant limiting factor that bounds the possible cancellable self-interference power. In this paper, a self-interference cancellation scheme for full-duplex orthogonal frequency division multiplexing systems is proposed. The proposed scheme increases the amount of cancellable self-interference power by suppressing the distortion caused by the transmitter and receiver nonlinearities. An iterative technique is used to jointly estimate the self-interference channel and the nonlinearity coefficients required to suppress the distortion signal. The performance is numerically investigated showing that the proposed scheme achieves a performance that is less than 0.5dB off the performance of a linear full-duplex system.
I. INTRODUCTION
In full-duplex systems, strong self-interference and radio-circuit nonlinearity limit how much interference can be cancelled. The paper proposes digital-domain nonlinear-distortion suppression for full-duplex OFDM and reports performance within 0.5 dB of a linear full-duplex system.
I. INTRODUCTION
- Strong self-interference and radio-circuit impairments prevent complete self-interference elimination in current full-duplex systems.Existing schemes that ignore nonlinearity are limited in cancellable self-interference power by the distortion level.Source: §I · p. 1
- A digital-domain cancellation scheme suppresses distortion from transmitter and receiver nonlinearities to increase cancellable self-interference power.The paper considers full-duplex OFDM systems in the presence of transmitter and receiver nonlinearities, oscillator phase noise, receiver Gaussian noise, and ADC effects.Source: §I · p. 1
- After three to four iterations, the proposed scheme achieves performance less than 0.5 dB off that of a linear full-duplex system.The iterations significantly suppress nonlinear distortion.Source: §I · p. 1
- The method jointly estimates the self-interference channel and nonlinearity coefficients to suppress distortion that otherwise limits channel-estimation accuracy.The channel-estimation error is distortion-limited while the channel is being estimated.Source: §I · p. 1
II. SIGNAL MODEL
The signal model describes simultaneous full-duplex OFDM transmission with transmitter and receiver impairments. It focuses on third-order nonlinear distortion and identifies that distortion as the dominant modeled noise source.
II. SIGNAL MODEL
- The transceiver operates its transmitter and receiver simultaneously on the same carrier, with phase errors modeled at both oscillators.The transmit chain includes OFDM modulation, up-conversion, and a power amplifier; the receive chain includes an LNA, down-conversion, quantization, and Fourier transformation.Source: §II · p. 1, Figure 1 · p. 2
- The model limits transmitter and receiver nonlinearities to third order, since only odd polynomial orders contribute in-band and higher orders can be neglected.The passage also notes that coexistence can introduce fifth-, seventh-, and ninth-order harmonics, with the seventh and ninth described as much smaller and ignorable.Source: §II · p. 2
- At the ADC output, the received baseband signal includes the desired signal, self-interference, total nonlinear distortion, phase noise, quantization noise, and receiver Gaussian noise.The nonlinear distortion model ignores distortion associated with the signal of interest because its power is small compared with the self-interference signal.Source: §II · p. 2
- Nonlinear distortion is the main modeled noise source, followed by phase noise, receiver Gaussian noise, and quantization noise.The comparison is shown across different received self-interference signal strengths for the transceiver in.Source: Figure 2 · p. 2, §II · p. 2
III. SELF-INTERFERENCE CANCELLATION WITH DISTORTION SUPPRESSION
The scheme suppresses nonlinear distortion using estimates of the self-interference channel and nonlinearity coefficients. It iterates their estimation with distortion reconstruction and subtraction, then subtracts the self-interference and distortion from data symbols.
III. SELF-INTERFERENCE CANCELLATION WITH DISTORTION SUPPRESSION
- Distortion suppression requires both the self-interference channel and nonlinearity coefficients, so the scheme estimates these quantities jointly through iteration.The channel estimate and known self-interference signal are used to estimate the nonlinearity coefficients.Source: §III · p. 3
- Each iteration estimates the channel, estimates the nonlinearity coefficients, reconstructs the distortion, and subtracts that estimate from the received signal.The four steps are repeated for a number of iterations.Source: §III · p. 3
- After estimation, the scheme subtracts both the self-interference signal and estimated distortion from each data OFDM symbol.Source: §III · p. 3
A. Channel estimation
The channel-estimation procedure uses a DFT-based estimate and reduces error by retaining the CIR samples that contain channel information. Its iterative use requires channel-estimation error below the distortion power, and the cyclic-prefix length is suggested for choosing L.
A. Channel estimation
- The iterative technique provides no gain if channel-estimation mean square error is not below the distortion power.Otherwise, performance is limited by channel-estimation error.Source: A. Channel estimation · p. 3
- Keeping the first L least-squares CIR samples and setting the rest to zero reduces estimation error by a factor of L/N.N is the number of subcarriers per OFDM symbol.Source: A. Channel estimation · p. 3
- Choosing L equal to the cyclic-prefix length is suggested because practical cyclic prefixes exceed the channel length.Source: A. Channel estimation · p. 3
B. Nonlinearity coefficients estimation
Nonlinearity coefficients are estimated successively to avoid an ill-conditioned least-squares matrix, with coefficient order chosen to reduce residual distortion when prior knowledge is unavailable.
B. Nonlinearity coefficients estimation
- At the self-interference training symbol, the signal-of-interest is absent, allowing the received signal model to be simplified.Source: B. Nonlinearity coefficients estimation · p. 3
- A successive one-by-one coefficient estimator avoids inversion of the often ill-conditioned least-squares matrix.The procedure resembles successive interference cancellation.Source: B. Nonlinearity coefficients estimation · p. 3
- Iterating the successive coefficient estimates can improve their accuracy.Source: B. Nonlinearity coefficients estimation · p. 3
- When transmitter-versus-receiver nonlinearity strength is unknown, the algorithm starts with the coefficient yielding smaller residual distortion power.With prior knowledge, it starts with the coefficient corresponding to the stronger nonlinearity.Source: B. Nonlinearity coefficients estimation · p. 4
IV. SIMULATION RESULTS AND DISCUSSIONS
Across the simulations, distortion suppression brings residual interference, distortion, and noise near the next bottleneck and to within 0.5 dB of a linear receiver. In distortion-dominated conditions, full-duplex rates exceed half-duplex rates at high SNR.
IV. SIMULATION RESULTS AND DISCUSSIONS
- The evaluation uses WiFi 802.11n parameters, TGn indoor channels, and compares against both a linear full-duplex case and nonlinear full-duplex without distortion removal.Performance criteria are achievable rate and residual interference plus distortion plus noise power.Source: §IV · p. 4
- Less than 0.5 dB separates the proposed scheme from a linear receiver after it suppresses distortion to the next bottleneck level.Figure 4 examines transmitter and receiver distortion levels; performance is similar whichever side's distortion dominates.Source: Figure 4 · p. 4, §IV · p. 4
- When another noise component dominates nonlinear distortion, the proposed scheme matches the case without distortion suppression.This indicates no performance degradation at low distortion levels.Source: §IV · p. 4, Figure 5 · p. 5
- At high SNR, full-duplex achieves a better rate than half-duplex when nonlinear distortion dominates other noise components.Source: §IV · pp. 4–5, Figure 6 · p. 5
V. CONCLUSION
The proposed digital-domain full-duplex OFDM cancellation scheme suppresses transmitter and receiver nonlinear distortion to increase cancellable self-interference power.
V. CONCLUSION
- The scheme suppresses nonlinear distortion to the next significant noise component and performs less than 0.5 dB below a linear full-duplex system.Source: §V · p. 5