Frequency-as-Aperture Imaging

Updated 5 February 2026

Frequency-as-Aperture is a concept that transforms frequency into a virtual spatial aperture, assigning specific frequency bands to spatial sampling roles for improved resolution and artifact reduction.
By synthesizing dynamic apertures from frequency-diverse signals, FaA systems achieve measurable gains, such as up to 24% lateral resolution improvement and enhanced contrast in imaging applications.
FaA techniques enable hardware simplification through single RF-chain architectures and adaptive beamforming, benefiting modalities like ultrasound, mmWave sensing, and synthetic aperture radar.

Frequency-as-Aperture (FaA) is a paradigm in wave-based imaging and sensing in which frequency is explicitly leveraged as a control parameter for the spatial aperture of a system, effectively allowing frequency to function as a virtual spatial aperture. This approach unifies techniques across array imaging (e.g., ultrasound, radar), mmWave communications, and electromagnetic lens design by reallocating spatial sampling and focusing from the traditional physical aperture domain to the frequency domain. The FaA principle directly assigns specific frequencies—or frequency bands—to discrete or continuous spatial sampling roles, enabling improved spatial resolution, artifact suppression, and hardware simplification in diverse applications (Schiffner et al., 2021, Schiffner, 2024, Ho et al., 29 Jan 2026, Hammond et al., 2014, Borcea et al., 2015).

1. Mathematical Foundations of Frequency-as-Aperture

In classical array imaging, spatial resolution and grating-lobe suppression are determined by the width of the physical aperture and the element pitch. The Frequency-as-Aperture principle modifies this association by making aperture width a function of frequency, or by utilizing frequency diversity to synthesize virtual apertures.

Ultrasound and Array Imaging

For a uniform linear array, the traditional receive F-number at focal depth $z_{\mathrm{f}}$ is

$F = \frac{z_{\mathrm{f}}}{A(z_{\mathrm{f}})},$

where $A(z_{\mathrm{f}})$ is the physical aperture width. FaA introduces a frequency-dependent $F$ -number,

$F(f) = \left[ \frac{c}{f \cdot p} - \sin x_{\min} \right]^{-1},$

where $p$ is element pitch and $x_{\min}$ is the enforced minimum main-lobe/grating-lobe angle, so that at each frequency $f$ , the associated aperture width is $W(f) = z_{\mathrm{f}} / F(f)$ . At low frequencies (large $\lambda$ ), wider apertures are safe and exploited for high resolution; at high frequencies (small $\lambda$ ), the aperture is automatically reduced to avoid spatial undersampling and grating lobes (Schiffner et al., 2021, Schiffner, 2024).

mmWave Communication/Sensing

In frequency-scanned leaky-wave antennas (LWAs), beam angle $\theta$ becomes a deterministic function of frequency:

$\theta(f) = \theta_0 + \kappa (f - f_c),$

where $\kappa$ is the LWA dispersion parameter, enabling $M$ distinct frequency “looks” over the communication bandwidth $B = f_{\max} - f_{\min}$ , which together compose an effective virtual aperture (Ho et al., 29 Jan 2026).

Synthetic Aperture Radar and Data Fusion

In synthetic aperture systems, frequency sub-bands can be treated as "frequency apertures," analogous to spatial sub-apertures. The forward signal model for each sub-band (with slow-time segmentation) enables joint estimation of spatial and frequency-dependent reflectivities through multi-measurement $\ell_1$ -minimization, exploiting the orthogonality of waveforms across sub-bands (Borcea et al., 2015).

2. Physical Implementations and Architectures

FaA has been realized across several platforms and modalities.

Frequency-Dependent Dynamic Apertures

In pulse-echo ultrasound and line-by-line scanning, the receive (and transmit) aperture is dynamically set per frequency according to system constraints such as grating-lobe suppression and depth-of-field conservation. The receive subaperture at each frequency is selected via real-time or precomputed frequency-dependent windows within a spectrally resolved beamformer (Schiffner et al., 2021, Schiffner, 2024).

Virtual Array Sensing in Wireless Radio

A single-RF-chain frequency-modulated continuous-wave (FMCW) mmWave transceiver combined with two orthogonal LWAs uses discrete LO frequencies to “scan” two spatial dimensions. This achieves $M$ -point angular sampling and range discrimination without explicit array steering or multiple RF chains (Ho et al., 29 Jan 2026).

Frequency-Dependent Metasurface Optics

A zoned metamaterial lens built from miniaturized element frequency selective surfaces (MEFSS) realizes an optics analog: at each frequency, the lens’s focal length $\ell(f)$ is directly specified, and the local phase shift is engineered to achieve the desired focus after correcting for aperture truncation. An adjustable iris provides further "frequency-as-aperture" tuning at the hardware level (Hammond et al., 2014).

Implementation Domain	Mechanism	Reference
Ultrasound Imaging	$F(f)$ -controlled subaperture size	(Schiffner, 2024, Schiffner et al., 2021)
mmWave Sensing	Frequency-indexed steering via LWA	(Ho et al., 29 Jan 2026)
Metamaterial Lenses	Phase profile $\phi(r;f)$ for focal control	(Hammond et al., 2014)

3. Resolution, Artifact Suppression, and Optimality

In all FaA frameworks, frequency diversity is exploited to optimally trade off spatial resolution, suppression of spatial aliasing, and artifact minimization under system constraints.

Ultrasound: Empirically, using $F(f)$ allows the system to use the largest possible aperture at each band-limited frequency, maximizing lateral resolution ( $\mathrm{FWHM} \propto \lambda \cdot F(f)$ ) while suppressing grating lobes by limiting aperture at high frequencies. Lateral resolution gains up to 12.8% (wire targets) and contrast improvements of 3.2% have been reported for frequency-dependent F-number approaches versus fixed- $F$ (Schiffner et al., 2021). Extensions achieve up to 24% lateral FWHM gains and 14.1% uniformity improvements in line-by-line scanning (Schiffner, 2024).
mmWave ISAC: In the FaA-Single architecture, angular resolution is determined by the effective frequency-swept aperture, yielding $\Delta\theta \approx 0.9^\circ$ over a 6 GHz sweep with $M=128$ frequency points, outperforming MIMO-based systems in efficiency under cost and power constraints (Ho et al., 29 Jan 2026).
Synthetic Aperture Radar: The use of frequency as an aperture dimension allows spatial and frequency-dependent reflectivity tasking through $\ell_1$ /MMV minimization, improving recoverability of complex scenes and supporting joint spatial-frequential imaging (Borcea et al., 2015).

4. Comparative Benefits, Trade-offs, and Metrics

FaA offers a new axis of system optimization.

Hardware Efficiency: Single RF chain designs become feasible, reducing cost and complexity compared to MIMO arrays (Ho et al., 29 Jan 2026).
Beam Profile Optimization: Dynamic apertures per frequency maximize main-lobe sharpness at all depths while preserving artifact suppression.
Architectural Efficiency Metric: $n = (1/\Delta\theta)/(N_{\text{RF}}\cdot D_{\text{phys}})$ quantifies resolution per unit hardware, showing >16x gains for FaA-Single vs. 1T3R-MIMO under identical constraints (Ho et al., 29 Jan 2026).
Computational Costs: Pre-calculable frequency-aperture maps and FFT-based beamforming limit algorithmic overhead in array imaging (Schiffner et al., 2021).
Flexibility: Adjustable apertures (via iris/diaphragm in optics or dynamic beamformer windows in arrays) allow real-time system adaptation to changing acquisition goals (Hammond et al., 2014, Schiffner, 2024).

A plausible implication is that FaA approaches inherently favor broadband or frequency-agile platforms, and may lose advantage in static, far-field, or narrowband scenarios where frequency diversity is unavailable or underexploited.

5. Applications and Empirical Validation

FaA strategies have been applied and validated in multiple domains:

Ultrasound Imaging: Both fast plane-wave imaging and line-by-line scanning demonstrate quantifiable improvements in resolution, contrast, and artifact suppression using frequency-dependent aperture controls (Schiffner et al., 2021, Schiffner, 2024). Phantom studies confirm consistency and statistical significance over multiple imaging configurations.
mmWave Sensing for 6G: Integrated simultaneous localization and communication nodes leverage FaA to enable privacy-preserving, battery-powered, embeddable ISAC devices, with robust spatial fingerprinting for smart environments (Ho et al., 29 Jan 2026).
Electron Cyclotron Emission Diagnostics: Zoned metamaterial lenses using frequency-as-aperture design enable focus alignment with frequency-resolved plasma emission layers in tokamaks, with performance controlled through hardware optics and expressible with general phase profiles $\ell(f)$ (Hammond et al., 2014).
Synthetic Aperture Radar: Fusion of directional and frequency sub-apertures enhances scene reconstruction, especially for anisotropic or frequency-selective reflectors (Borcea et al., 2015).

6. Extensions, Limitations, and Future Directions

FaA is extendable to other wave modalities (photoacoustics, RF, optics) and higher-dimensional arrays. Time-domain implementations (e.g., band-specific filter banks for ultrasound) and software-defined frequency-aperture mappings (in mmWave ISAC) are suggested. Future research includes calibration methods for frequency-dependent antenna patterns, networked spatial fingerprint fusion in distributed sensor settings, and integration with adaptive beamformers for diffuse or dynamic scenes (Ho et al., 29 Jan 2026, Schiffner et al., 2021, Schiffner, 2024). In far-field or high-dynamic-rate settings, the sequential (frequency-indexed) sampling advantage of FaA may diminish.

In summary, Frequency-as-Aperture reconfigures the interplay between frequency and spatial sampling across a broad spectrum of wave-based imaging and sensing systems, yielding practical and theoretical advances in resolution, efficiency, and hardware simplification (Schiffner et al., 2021, Schiffner, 2024, Ho et al., 29 Jan 2026, Hammond et al., 2014, Borcea et al., 2015).