Digital Self-Interference Cancellation for In-Band Full-Duplex Systems Based on Transformer Neural Operators
This paper tackles digital-domain self-interference cancellation in in-band full-duplex radios, where conventional adaptive filters degrade when the desired signal is much stronger than the residual interference.
The authors reformulate cancellation as an operator-learning problem: a Transformer neural operator learns the temporal evolution of the self-interference channel from historical observations and reconstructs the current interference from the known transmit signal.
Trained in a desired-signal-free setting and applied during contaminated reception, the method reduces reliance on corrupted instantaneous updates and reports strong high-SIR stability.
The paper presents a novel operator-learning perspective on digital self-interference cancellation and delivers substantial empirical high-SIR stability. However, the theoretical contribution is limited to an elementary continuity observation, and the experimental validation relies on overly simplistic synthetic channels that do not exercise the full complexity of the proposed architecture.
The clean-training/contaminated-inference formulation is a compelling departure from conventional iterative adaptive filtering, and the dynamic-channel experiments demonstrate that the method remains stable when the desired signal and channel vary independently. The reported gains are notable: the method holds its desired-signal-free performance even at 100 dB SIR, whereas iterative baselines collapse.
Theorem 1 is a direct consequence of assuming continuity: it provides no rates, no architecture-dependent constants, and no guarantee beyond the generic bound in (27). The static-channel experiment relies on the trivial scalar channel $l(n)=0.001x(n)$, so the reported pairwise difference of no greater than 0.05 dB across all tested SIR values is uninformative for realistic multipath or nonlinear channels. Moreover, the dynamic setup undergoes abrupt changes every 500 samples but is recalibrated only every 20,000 samples, leaving behavior during the many intervening transients largely unreported.
The evidence supports the main high-SIR claim: at 30 dB SIR the proposed method improves NMSE by 5–15 dB over baselines that were given the advantage of retrospective grid search over the full trajectory. However, the paper does not benchmark against the closely related in-context regression work of Balaji et al. [36] under a shared protocol, and the severe 'opposite cancellation' suffered by baselines at 100 dB SIR suggests they were pushed outside their valid operating regimes rather than fairly tuned for such extremes.
Reproduction is severely hindered by the unavailability of both data and code. The synthetic channel generation is lightly specified, the abrupt-change dynamic model is described only qualitatively, and key deployment metrics such as parameter count, inference latency, and random seeds are omitted. While hyperparameters and learning-rate schedules are reported, the absence of publicly shareable data or implementation details blocks independent validation.
Digital-domain self-interference cancellation is essential in in-band full-duplex radios after passive suppression and analog cancellation. It is particularly challenging when the desired signal is strong relative to the residual self-interference. In this regime, conventional iterative digital cancellers update their coefficients from desired-signal-contaminated received samples and may suffer substantial performance degradation. This paper reformulates digital-domain self-interference cancellation as an operator-learning problem. The proposed framework learns the temporal evolution of the residual self-interference channel from historical observations and uses the learned operator together with the known local transmit signal to reconstruct the current residual self-interference. A Transformer neural operator is used to realize this framework. The resulting method is trained in a desired-signal-free setting and then applied in the presence of a desired signal. A continuity-based analysis characterizes the sensitivity of the learned operator to the resulting input perturbation. Numerical experiments in static and time-varying channels, including time-varying SIR conditions, demonstrate stable cancellation performance over a wide range of interference levels. At 30 dB SIR, the proposed method improves the cancellation performance by 5--15 dB over the considered conventional methods.
Pick a starting point or write your own. Challenges run in the background, so you can keep reading while the AI investigates.
No challenges yet. Disagree with the review? Ask the AI to revisit a specific claim.