Software Defined Physical Modulation for Adaptive Sensing

19 concepts11 papers

Traditional sensing systems are bound by their physical geometry and material properties. A standard camera relies on a fixed glass lens and a static color filter array; a conventional radio receiver uses a hardwired filter bank. If you want to change the camera's field of view, you must mechanically move the lens elements. If you want to isolate a different wireless band, you must switch physical circuits.

The core idea of software-defined physical modulation is to replace these static elements with programmable, active modulators. By inserting a software-controlled layer—such as a liquid crystal display (LCD), a spatial light modulator (SLM), or a digital multiplexer array—directly into the path of an incoming wave (light or radio), we can dynamically reshape the system's physical response on the fly. This shifts the burden of adaptation from rigid mechanical hardware to flexible software algorithms, allowing a single device to fundamentally alter its function from one millisecond to the next.

Reshaping Light Paths Without Moving Parts

How do you change a camera's perspective, focus, or contrast modality without moving a single lens? The key is to control the aperture and illumination dynamically. Instead of using a mechanical iris or physical optical inserts, we can use a programmable control surface to block or pass light selectively.

Zomet et al. demonstrated this by replacing traditional refractive lenses entirely with a stack of programmable, light-attenuating LCD layers. By changing the transparency patterns on these layers in real time, the authors created a software-defined aperture. This architecture enables instantaneous digital panning, tilting, and field-of-view adjustments without any moving parts.

Static Optics:      [Incoming Light] ---> [Fixed Glass Lens] ---------> [Sensor]
Software-Defined:   [Incoming Light] ---> [Programmable LCD Stack] ---> [Sensor]
                                                 ^
                                         (Software Control)

This concept of geometric flexibility is crucial because the optimal aperture pattern depends heavily on the specific task at hand. For instance, depth estimation and defocus deblurring require entirely different optical profiles. To address this, Zhou et al. utilized a Liquid Crystal on Silicon (LCoS) device to implement dynamic coded apertures, allowing the camera to optimize its mask frame-by-frame based on changing scene conditions. As noted by Ye et al., transitioning from static physical slits to programmable apertures allows real-time optimization of camera geometry to fit the specific structural layout of a scene.

This software-defined approach applies equally to how we illuminate a scene. In microscopy, switching between brightfield, darkfield, and phase-contrast imaging traditionally requires swapping physical optical inserts beneath the sample. Phillips et al. replaced these mechanical inserts with a programmable LED array. By using software to electronically toggle specific patterns on the LED dome, the system alters the angle of illumination instantaneously, replicating the performance of high-end laboratory microscopes on a portable device.

Optical Computation and Spectral Coding

Beyond spatial manipulation, programmable modulators can fundamentally alter how a system perceives color and wavelengths. Standard cameras reduce the rich, high-dimensional spectrum of light into just three fixed channels: Red, Green, and Blue. Software-defined spectral modulation allows us to slice and encode the spectrum arbitrarily.

To achieve this, Mohan et al. introduced an agile spectrum imaging system. By placing a diffraction grating and a programmable mask in the optical path, they dispersed incoming light into its constituent wavelengths across a spatial light modulator. By selectively attenuating specific columns of light before recombining them onto the sensor, the hardware can implement arbitrary spectral transmission functions. This allows the camera to dynamically adapt its response to detect metameric materials or eliminate glare.

[Light] ---> [Diffraction Grating] ---> [Spatial Light Modulator] ---> [Recombined] ---> [Sensor]
                   (Disperses Wavelengths)       (Programmable Mask)

When paired with machine learning, this architecture can perform physical computation before the light even reaches a digital sensor. Saragadam et al. built a programmable spectrometer that uses a coded aperture and an SLM to implement learned spectral filters (such as those from a neural network or support vector machine). Instead of capturing a massive hyperspectral data cube and processing it on a computer, the hardware physically filters the light to perform per-pixel material classification at video rates, dramatically reducing data overhead and power consumption.

This active encoding is also central to compressive sensing. Yuan et al. demonstrated that replacing fixed dispersive optics with programmable spatial light modulators enables software-defined spectral coding for hyperspectral imaging. By dynamically tuning the sensing patterns to match environmental conditions, the camera optimizes data capture on the fly.

Scaling to Computing and Radio Frequencies

The core intuition of software-defined physical modulation—using a programmable layer to manipulate waves—extends naturally beyond imaging into optical computing and radio frequency (RF) signal processing.

In optical computing, cascaded spatial light modulators can serve as programmable diffractive layers that manipulate the phase of light to perform mathematical operations. Tang et al. leveraged this mechanism to expand Diffractive Optical Neural Networks (DONNs) into scientific computing. Because the phase maps on the SLMs can be updated instantaneously via electrical signals, the same physical hardware can switch between entirely different computational tasks without needing to be physically re-fabricated. Similarly, Nakamura et al. integrated programmable liquid crystal filters into light-field acquisition pipelines, transforming the hardware into a software-defined measurement system that can dynamically switch between multiplexed and adaptive sensing strategies depending on the light source.

This paradigm shifts seamlessly from light waves to electromagnetic waves in communication. In software-defined radio (SDR), handling multiple wireless standards simultaneously typically requires complex, power-hungry hardware updates to accommodate different channel bandwidths. Hao et al. solved this by integrating a multiplexer array into a fixed cascaded filter bank. By simply toggling control bits in software, the system aggregates fine-grained sub-channels to dynamically allocate non-uniform bandwidths without modifying the underlying physical circuit.

Fixed Filter Bank ---> [ Multiplexer Array ] ---> Dynamic Bandwidth Output
                              ^
                      (Software Bit Switch)

In the optical-microwave domain, Li et al. demonstrated a reconfigurable microwave photonic filter based on optical microcombs. By programmatically updating the power levels of individual optical taps, a single physical architecture can be instantly repurposed to switch between low-pass, high-pass, and multi-band configurations without requiring structural changes to the optical or electrical paths.

Whether manipulating the phase of a light wave for neural network inference, shifting virtual apertures for a lensless camera, or toggling multiplexer bits for a radio channelizer, the underlying mechanism remains identical. By embedding a programmable control surface directly into the physical wave-propagation path, software-defined hardware bridges the gap between the analog world and digital computation, allowing physical sensors to adapt to their environments in real time.

Key Foundational Work

Lensless Imaging with a Controllable Aperture (2006) — contributed 2 concept elements to this direction.
Agile Spectrum Imaging: Programmable Wavelength Modulation for Cameras and Projectors (2008) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_programmable_spectral_modulation_a21216f6,
  title = {programmable spectral modulation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/programmable-spectral-modulation/rdr_a21216f62fce4e75ae1976d7873810fa},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_programmable_spectral_modulation_a21216f6_v1,
  title = {programmable spectral modulation},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/programmable-spectral-modulation/rdr_a21216f62fce4e75ae1976d7873810fa/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}

See something wrong? Report a mistake