Self-Interference Cancellation in Full-Duplex Systems

475 concepts309 papers2 proposals

Imagine trying to hear a friend whispering from across a crowded room while you are screaming at the top of your lungs. Your own voice completely overwhelms your ears, making it impossible to catch the whisper.

In wireless communications, this is the fundamental challenge of in-band full-duplex operation—the ability to transmit and receive data simultaneously on the exact same frequency band. When a radio transmits a high-power signal, that signal leaks directly into its own co-located receiver. This leakage, known as Self-Interference (SI), can be 50 to 110 dB stronger than the incoming signal you actually want to hear. As noted by Ian P. Roberts et al. in Millimeter-Wave Full Duplex Radios: New Challenges and Techniques, this power disparity can be billions of times greater than the desired signal, effectively deafening the receiver hardware.

If solved, full-duplex systems can theoretically double spectral efficiency and unlock advanced capabilities like monostatic radar sensing on standard communication hardware. Achieving this requires understanding why self-interference is so destructive and how modern systems cancel it across multiple domains.

The Hardware Bottleneck: Receiver Saturation

A common misconception is that self-interference can be easily subtracted digitally since the device knows exactly what signal it just transmitted. However, this ignores a critical physical constraint: the limited dynamic range of receiver hardware.

Before a wireless signal can be processed digitally, it must pass through analog components like Low-Noise Amplifiers (LNAs) and Analog-to-Digital Converters (ADCs). These components have strict upper limits on the signal power they can handle. When the massive power of the self-interference signal hits the receiver front-end, it saturates these components.

[ Massive SI + Weak Echo ] ---> [ LNA / ADC ] ---> [ Saturated / Clipped Digital Output ]
                               (Saturates!)        (Information is permanently lost)

Once an ADC saturates, the signal is "clipped," and the subtle variations of the weak, incoming signal are permanently lost to quantization noise. As established by Brian P. Day et al. in Full-Duplex MIMO Relaying: Achievable Rates under Limited Dynamic Range, this hardware saturation prevents the system from accurately recovering the desired signal regardless of how sophisticated the downstream digital cancellation algorithms are. Therefore, self-interference must be mitigated before it passes through the receiver's main amplification and digitization stages.

The Multi-Stage Cancellation Pipeline

To prevent receiver saturation, practical full-duplex systems employ a multi-stage cancellation pipeline that chips away at the self-interference power across the spatial, analog, and digital domains.

+-----------------------------------------------------------------------+
|                         Self-Interference (SI)                        |
+-----------------------------------------------------------------------+
                                    |
                                    v
                       +-------------------------+
                       |     Spatial Domain      |  <-- Beamforming & Nulling
                       +-------------------------+
                                    |
                                    v
                       +-------------------------+
                       |      Analog Domain      |  <-- RF Injection & FDE
                       +-------------------------+
                                    |
                                    v
                       +-------------------------+
                       |     Digital Domain      |  <-- Non-linear LMS Algorithms
                       +-------------------------+
                                    |
                                    v
                        [ Clean Received Signal ]

1. Spatial Domain Cancellation

The first line of defense uses multi-antenna arrays to physically isolate the transmitted signal from the receiver. By leveraging beamforming, the transmitter can shape its radiation pattern to steer energy away from its own receive antennas, creating a "spatial null."

In wideband millimeter-wave (mmWave) systems, this becomes more complex because the self-interference channel varies across different frequencies. To address this, Ian P. Roberts et al. developed a hybrid analog/digital architecture in Frequency-Selective Beamforming Cancellation Design for Millimeter-Wave Full-Duplex that accounts for frequency-selective wideband channels, using spatial steering to protect the receiver front-end from saturation.

2. Analog and RF Cancellation

Even with spatial nulling, substantial energy still leaks into the receiver. The next stage occurs in the RF domain, before the ADC. The system takes a reference copy of the transmitted signal, adjusts its amplitude and phase to match the leakage path, and injects this inverted signal directly into the receive path to destructively interfere with the SI.

To make this practical on compact hardware, Jin Zhou demonstrated in Demo abstract: Full-duplex with a compact frequency domain equalization-based RF canceller that a frequency-domain equalization (FDE) RF canceller can achieve up to 95 dB of total self-interference cancellation across a 20 MHz bandwidth, proving that analog cancellation can be implemented efficiently without massive, expensive components.

3. Digital Cancellation

Once the self-interference has been suppressed enough to prevent ADC saturation, the remaining residual interference is digitized. The final cleanup happens in software.

Real-world hardware introduces non-linear distortions (such as power amplifier imperfections), meaning the leaked signal is no longer a perfect copy of the original transmission. To solve this, Yili Xia et al. proposed an adaptive filtering approach in An Augmented Nonlinear LMS for Digital Self-Interference Cancellation in Full-Duplex Direct-Conversion Transceivers. By using an augmented non-linear Least Mean Squares (LMS) algorithm, the digital processor models and subtracts these complex hardware distortions, isolating the true external signal.

Why It Matters: ISAC and Backscatter

Perfecting self-interference cancellation is not just about faster downloads; it is a foundational enabler for next-generation wireless paradigms like Integrated Sensing and Communication (ISAC) and Backscatter Communication.

Monostatic Sensing over Wi-Fi and 6G

In 6G networks, base stations are envisioned to act simultaneously as communication transmitters and radar sensors. In a monostatic radar setup, a single device transmits a signal and listens for the faint environmental echoes bouncing off objects (like cars or pedestrians).

Without advanced SIC, the high-power communication transmission completely blinds the radar receiver, masking these weak echoes. Shu Cai et al. demonstrated in Symbol-Level Precoding-Based Self-Interference Cancellation for ISAC Systems that optimizing transmission waveforms at the symbol level can drastically reduce residual interference, maintaining high target detection accuracy.

Showing the practical consumer applications of this concept, Chao Hu et al. developed ISAC-Fi: Enabling Full-fledged Monostatic Sensing over Wi-Fi Communication, which implements a real-time SIC scheme on standard Wi-Fi devices. This allows a single router to perform high-precision human imaging and motion tracking without interrupting normal data throughput.

Extracting Weak Backscatter Signals

A similar power differential problem occurs in ambient backscatter communication, where ultra-low-power IoT tags communicate by passively reflecting signals from existing RF sources (like TV towers or Wi-Fi routers).

[ Ambient Source ] ------------------------> [ Receiver ] (Massive Direct-Path Interference)
        \                                         ^
         \--> [ Passive Tag ] --(Weak Reflection)-/

As explained by Takanori Hara et al. in Ambient OFDM Pilot-Aided Backscatter Communications: Concept and Design, the direct-path signal from the ambient source to the receiver acts as massive interference that overwhelms the incredibly weak reflection from the passive tag. To extract the tag's data, the receiver must treat the strong direct-path signal similarly to self-interference.

By decoupling transmitters and receivers or using advanced signal separation, systems like the one proposed by Yanwen Wang et al. in Toward a Low-Cost Software-Defined UHF RFID System for Distributed Parallel Sensing manage this direct-link leakage to enable parallel, low-cost environmental sensing.

Ultimately, whether a device is trying to listen to a distant cell phone, a radar echo, or a passive RFID tag, managing the overwhelming energy of its own local environment relies on the same core intuition: suppressing the local giant so you can hear the distant whisper.

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 313 papers
  2. Version 1 · 2026-10-04 · 234 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_full_duplex_self_interference_c955eec1,
  title = {full-duplex self-interference},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/full-duplex-self-interference/rdr_c955eec19e3a45e39d84aacf32b325d6},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 2, frozen text):

@misc{lacuna_full_duplex_self_interference_c955eec1_v2,
  title = {full-duplex self-interference},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/full-duplex-self-interference/rdr_c955eec19e3a45e39d84aacf32b325d6/v/2},
  note = {Version 2, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake