Source-linked AI summary
Passive Self-Interference Suppression for Full-Duplex Infrastructure Nodes
Evan Everett, Achaleshwar Sahai, Ashutosh Sabharwal
TL;DR
Passive suppression accounts for a substantial share of self-interference reduction, yet its capabilities and bottlenecks were not well characterized. This paper measures three passive mechanisms across antenna configurations and environments, finding over 70 dB suppression in low-reflection settings while identifying reflection and frequency-selectivity limits.
Problem
Passive suppression contributes substantially to total self-interference reduction, but its capabilities and fundamental bottlenecks were not known.
Method
The paper uses network-analyzer measurements to characterize directional isolation, absorptive shielding, and cross-polarization across antenna configurations in reflective and nonreflective environments.
Results
More than 70 dB of passive self-interference suppression is achieved in anechoic conditions when the three mechanisms are combined, while reflected paths limit suppression in reflective environments.
Takeaways & Limitations
Full-duplex infrastructure designs should account for reflected paths and the increased frequency selectivity of residual self-interference after passive suppression.
Takeaways & Limitations
Passive suppression is fundamentally limited by environmental reflections, which the mechanisms cannot predictably attenuate.
Abstract
from arXiv · showhide
Recent research results have demonstrated the feasibility of full-duplex wireless communication for short-range links. Although the focus of the previous works has been active cancellation of the self-interference signal, a majority of the overall self-interference suppression is often due to passive suppression, i.e., isolation of the transmit and receive antennas. We present a measurement-based study of the capabilities and limitations of three key mechanisms for passive self-interference suppression: directional isolation, absorptive shielding, and cross-polarization. The study demonstrates that more than 70 dB of passive suppression can be achieved in certain environments, but also establishes two results on the limitations of passive suppression: (1) environmental reflections limit the amount of passive suppression that can be achieved, and (2) passive suppression, in general, increases the frequency selectivity of the residual self-interference signal. These results suggest two design implications: (1) deployments of full-duplex infrastructure nodes should minimize near-antenna reflectors, and (2) active cancellation in concatenation with passive suppression should employ higher-order filters or per-subcarrier cancellation.
I. INTRODUCTION
This paper presents a measurement-based characterization of passive self-interference suppression for full-duplex infrastructure nodes. It examines three mechanisms, their limitations, and implications for combining passive and active cancellation.
- Motivation: Passive suppression isolates transmit and receive antennas and accounts for a large portion of suppression in existing full-duplex designs.One cited design achieved 85 dB total suppression, including 65 dB passive and 20 dB active suppression.
- Contribution: The study characterizes directional isolation, absorptive shielding, and cross-polarization, including their combinations and effects on the self-interference channel.Its primary contribution is experimental characterization of passive suppression methods for full-duplex infrastructure nodes.
- Main results: Environmental reflections limit passive suppression because the mechanisms attenuate the direct path but cannot predictably attenuate reflected self-interference paths.The authors therefore recommend minimizing nearby reflectors in full-duplex infrastructure deployments.
- Main results: Passive suppression generally decreases the coherence bandwidth of the residual self-interference channel by leaving reflected paths relatively unaffected.Active cancellation should therefore handle a frequency-selective residual signal, for example with per-subcarrier cancellation.
- Prototype: A prototype combining the studied passive mechanisms with existing active cancellation achieved more than 90 dB total self-interference suppression and outperformed a comparable half-duplex link beyond 100 m.The prototype was evaluated in an outdoor full-duplex uplink using the WARPLab framework.
- Implications: Passive suppression is effective primarily against the direct path, while residual multipath can be addressed with frequency-selective active cancellation.Both passive and active suppression also encounter limitations, including reflected paths, phase noise, and limited dynamic range.
II. EXPERIMENT DESIGN
The experiment evaluates passive suppression across five antenna architectures using directional isolation, absorptive shielding, and cross-polarization. Measurements vary antenna orientation, separation, polarization, and shielding conditions.
- Antenna configurations: Five antenna configurations compare directional and omnidirectional architectures with different beam orientations and antenna separations.Directional configurations use 90° beamwidth antennas with 90° or 60° beam separation and 50 cm or 35 cm spacing; omnidirectional configurations provide comparison cases.
- Antenna configurations: Beamwidth denotes the 3 dB angular span over which antenna gain remains within 3 dB of its maximum.The illustrated conical patterns are figurative; actual patterns include gradual gain rolloff and sidelobes.
- Antenna configurations: Configuration II uses 90° beamwidth antennas with 60° beam separation, producing 30° overlap between coverage zones.Configurations I and II use 50 cm separation, while Configuration III reduces separation to 35 cm.
- Absorptive shielding: Absorptive shielding places a slab of broadband RF absorber between the antennas to attenuate self-interference.The AN-79 absorber used in the experiments is a 4.25-inch polyurethane-foam slab with a carbon gradient, rated by its manufacturer for up to 25 dB absorption.
- Cross-polarization: Directional antennas support dual polarization, allowing measurements of both co-polarized and cross-polarized coupling.The omnidirectional antennas support only vertical polarization, so cross-polarization was not studied in omnidirectional configurations.
B. Environments
The study compares passive suppression in a low-reflection anechoic chamber and a highly reflective room, while quantifying residual-channel frequency selectivity from measured responses.
- Measurement environments: Measurements were first conducted in a shielded anechoic chamber and then repeated in a highly reflective metal-walled room.The chamber minimizes reflections and external interference; the reflective room represents an environment with substantial environmental reflections.
- Suppression notation: Passive-mechanism combinations are represented by mech, with ∅ for no suppression, D for directional isolation, A for absorptive shielding, and C for cross-polarization.Antenna configuration constrains the combinations that can be evaluated.
- Suppression metric: Passive suppression is defined as the ratio of transmit power to self-interference power incident on the receiver, equivalently the inverse power of the measured frequency response.The quantity is computed from the measured frequency response for each mechanism combination.
- Frequency selectivity: Coherence bandwidth is the frequency range over which the channel response remains at least 90% correlated.The study estimates it from the RMS delay spread obtained from the measured power-delay profile.
- Frequency selectivity: The normalized power-delay profile provides the distribution used to compute mean delay and RMS delay spread.The RMS delay spread is then used to approximate coherence bandwidth from the empirical channel response.
III. EXPERIMENT RESULTS
Measurements compare passive suppression across anechoic and reflective environments using directional isolation, absorptive shielding, and cross-polarization. The results show strong suppression in low-reflection conditions but substantially weaker gains when environmental reflections dominate.
- Passive suppression measurements: 45 dB of passive suppression is measured for Configuration I without absorptive shielding or cross-polarization in the anechoic chamber.The configuration combines directional isolation with the antenna arrangement used in the low-reflection environment.
- Passive suppression measurements: More than 70 dB of passive suppression is achieved with directional isolation, absorptive shielding, and cross-polarization together in the anechoic chamber.This result holds across all three directional-antenna configurations.
- Environmental reflections: Environmental reflections limit passive self-interference suppression because the mechanisms suppress direct paths more effectively than reflected paths.After the direct path is reduced, reflected self-interference can become the dominant residual component.
- Environmental reflections: 10 dB of additional suppression comes from absorptive shielding in the anechoic chamber, compared with less than 3 dB in the reflective room.The smaller reflective-room improvement indicates that the shielding benefit depends strongly on the environment.
- Environmental reflections: Passive suppression mechanisms can make reflected-path tails dominant after suppressing the direct path, explaining weak incremental gains in reflective environments.Time-domain measurements are used to test this hypothesis across different mechanism combinations.
- Design implication: Full-duplex infrastructure nodes should be deployed as far as possible from potential reflectors.Where reflective environments cannot be avoided, passive suppression should not be pursued beyond the reflection strength.
B. Impact of Passive Suppression Frequency Selectivity of Self-interference Channel
Passive suppression generally makes the residual self-interference channel more frequency-selective by suppressing the direct path while leaving reflected paths relatively intact.
- Passive self-interference suppression generally decreases coherence bandwidth, increasing frequency selectivity in the residual channel.Measurements illustrate greater channel-gain variation with frequency as suppression mechanisms are applied.
- Figure 5 compares self-interference frequency responses, while Figure 6 illustrates direct and reflected paths underlying the coherence-bandwidth trend.Figure 6’s example has PD = 25 dB, PR = 5 dB, DRR = 20 dB, and TR = 20 s.
- Suppressing the direct path leaves a residual channel formed by many reflected paths, whose constructive and destructive combination produces frequency selectivity.Passive suppression transforms a line-of-sight-dominated channel into a low-power multipath channel.
- Analytical example: As direct-path power decreases relative to reflected-path power, the power-delay profile develops a heavier tail, increasing RMS delay spread and reducing coherence bandwidth.The analytical model relates this trend to decreasing direct-to-reflected ratio, DRR.
- Analytical example: With no passive suppression, the direct path dominates and the channel is relatively frequency-flat; stronger suppression lowers coherence bandwidth until it levels off.The limiting behavior occurs when the direct path is effectively eliminated and delay spread converges to the environment’s reverberation spread.
2) Coherence Bandwidth Measurements:
Measurements show that passive suppression reduces coherence bandwidth by suppressing the direct path while leaving reflected paths relatively unaffected. The effect is stronger in reflective environments and increases the complexity required for active cancellation.
- Coherence bandwidth measurements: Passive suppression increases direct-path suppression while reducing the residual channel’s coherence bandwidth.This occurs because reflected paths remain, producing a heavier-tailed power delay profile and greater frequency selectivity.
- Anechoic chamber: In the anechoic chamber, coherence bandwidth is 10–25 MHz without suppression and 2–15 MHz with 50–70 dB suppression.Combining cross-polarization and absorptive shielding produces more than 70 dB suppression, after which coherence bandwidth levels off around 3 MHz.
- Reflective room: In the reflective room, coherence bandwidth changes more abruptly because the reflection floor is approximately −45 dB.With one or more suppression mechanisms, the direct path falls below the reflected-path level and coherence bandwidth converges toward the reflection-limited regime.
- Design implications: Active cancellation may require more filter taps as passive suppression strengthens, especially in high-reflection environments.For 20 MHz signals, low-reflection measurements require 1–2 taps at 25–45 dB suppression and 10–15 taps at 60–80 dB; high-reflection measurements require 7–12 taps initially and nearly 30 above 40 dB.
- Design implications: The measurements support combining passive and active suppression, while the relative efficiency of the two approaches remains unresolved.Prior results summarized in the section indicate that neither approach alone outperforms their concatenation, but their gains are not dB-additive.
B. Impact of Direct-path Suppression on Capacity
The capacity analysis models residual self-interference as colored Gaussian noise after cancellation of the known transmit signal. Numerically, capacity stops increasing once direct-path strength reaches the reflection level, motivating deployment designs that reduce nearby reflections.
- Capacity analysis: Theorem 1 gives the full-duplex uplink capacity for the modeled self-interference channel.Known self-interference from the intended transmit signal is canceled, leaving transmitter noise filtered by the self-interference channel and receiver noise.
- Capacity analysis: Capacity is obtained by applying spectral water-filling to the resulting colored Gaussian noise channel.The noise power spectral density is SN(f) = |√NT HI(f)+√NR|2, with expectations taken over the channel responses.
- Capacity results: The capacity comparison includes half-duplex operation with equal uplink/downlink time sharing and ideal full-duplex operation with zero self-interference.Half-duplex includes a 1/2 pre-log factor and doubles user transmit power during its transmission interval.
- Capacity results: Once direct-path strength PD is suppressed to the reflection level PR, capacity no longer grows.The numerical evaluation uses WiFi-typical parameters, including PT = 0 dBm, NT = −30 dBm, NR = −90 dBm, PS = −60 dB, and TR = 32.
C. Influence of Increased Frequency Selectivity on Active Cancellation
The analysis studies active cancellation when the canceler has a limited number of taps. As passive suppression increases frequency selectivity, insufficient tap length causes cancellation performance to degrade.
- Limited-tap cancellation: The limited-tap analysis constrains the canceler to NTap ≤ TR taps and subtracts the known transmit signal filtered by the first NTap channel taps.The model assumes perfect knowledge of the retained channel segment.
- Cancellation metrics: Passive suppression αP and active cancellation αA are defined as average power ratios before and after suppression or cancellation.The expectations are taken over the specified channel-gain distributions.
- Numerical evaluation: For NTap = 1, 24, and 32, numerical evaluation uses PR = −60 dB, TR = 32, and the system parameters from Figure 9(a).These settings evaluate how cancellation changes as the available filter length approaches the reflection extent.
- Numerical evaluation: A single-tap canceler performs well at low passive suppression but degrades as stronger suppression makes the channel more frequency selective.The degradation reflects the inability of a short filter to represent the increasingly frequency-selective residual channel.
V. FULL-DUPLEX PROTOTYPE WITH COMBINED PASSIVE AND ACTIVE SUPPRESSION
A measured outdoor prototype combines passive mechanisms with per-subcarrier analog and digital cancellation to evaluate full-duplex uplink performance against half-duplex. Cross-polarization adds suppression and improves rate gains across substantial path losses.
- Prototype evaluation: Per-subcarrier analog and digital cancellation mechanisms were used to combine active cancellation with passive suppression.Analog cancellation injects a wideband cancellation waveform before the RF front end; digital cancellation follows channel estimation.
- Prototype evaluation: 60 dB of passive suppression is achieved on average with directional isolation and absorptive shielding.The prototype used 100 packets transmitted at 7 dBm and measured suppression through RSSI.
- Prototype evaluation: 86% improvement over half-duplex occurs at 86 dB path loss with directional isolation, absorptive shielding, and active cancellation.
- Prototype evaluation: 96% gain over half-duplex at 86 dB path loss with cross-polarization nearly doubles the rate.The added cross-polarization provides 10 dB of total suppression and shifts equivalent gains to 10 dB more path loss.
- Prototype evaluation: The prototype full-duplex design outperforms half-duplex beyond effective ranges exceeding 100 m without cross-polarization and up to 200 m with it.Rates were evaluated while path loss varied across mobile locations corresponding to 50–200 m ranges.
VI. DISCUSSION: IMPACT ON THE USER’S ACCESS TO THE INFRASTRUCTURE NODE
The discussion distinguishes two-node and three-node full-duplex infrastructure scenarios according to whether the user device must support full-duplex operation. The scenarios differ in their hardware requirements and link roles.
- Access scenarios: Three-node full-duplex does not require full-duplex-capable user devices, unlike two-node full-duplex.The paper expects three-node scenarios to be among the first applications because full-duplex hardware burdens mobile devices.
- Access scenarios: Two-node full-duplex has the infrastructure node simultaneously transmit and receive to a single full-duplex user.
- Access scenarios: Three-node full-duplex has the infrastructure node receive from one user while transmitting to another user.
A. Impact of Directional Isolation
Directional isolation is discussed alongside absorptive shielding and cross-polarization as passive mechanisms whose effectiveness and user implications depend on deployment conditions. The paper emphasizes reflected paths and polarization environment as practical boundaries.
- Deployment implications: Directional isolation is often better suited to three-node than two-node full-duplex applications when antenna coverage areas are disjoint.Configuration I cannot transmit and receive in the same direction, limiting its suitability for two-node transmissions.
- Deployment implications: Directional antenna systems require control to service each user with the antenna or antennas having the strongest gain in that user’s direction.The paper notes that adaptive antenna selection and dynamic user selection are existing approaches.
- Cross-polarization: In non-line-of-sight environments, scattering de-polarizes signals, making users relatively agnostic to the infrastructure node’s polarization state.A single-antenna user with arbitrary polarization can receive and transmit through the infrastructure node’s orthogonally polarized antennas.
- Cross-polarization: In line-of-sight environments, cross-polarization can reduce a single-polarized user’s SNR unless the user matches or tracks the infrastructure node’s polarization.Dual-polarized antennas and orientation tracking are discussed as ways to address this requirement.
- Limitations: Reflected paths limit passive suppression, while removing the dominant direct path can make residual self-interference more frequency selective.These limitations motivate designs that account for both passive suppression capabilities and their channel effects.
APPENDIX A MEASUREMENT DETAILS AND UNCERTAINTY ANALYSIS
The appendix describes network-analyzer measurement procedures and uncertainty analysis for the self-interference channel. Calibration, bandwidth, averaging, and setup parameters determine how measurement uncertainty is assessed.
- Measurement setup: A two-port network analyzer measures S-parameters by sweeping continuous-wave inputs and measuring output amplitude and phase.
- Measurement setup: Calibration after the cables and connectors places the phase and amplitude reference directly at the intended antenna ports.
- Measurement setup: A 20 kHz intermediate-frequency bandwidth balances noise floor and frequency sweep time, while each frequency point averages 20 repeated measurements.
- Uncertainty analysis: The uncertainty calculator accounts for thermal noise, averaging, and calibration imperfections rather than only empirical variation across repeated measurements.The relevant setup parameters include the 20 kHz bandwidth and averaging factor of 20.