ABSense: Sensing Electromagnetic Waves on Metasurfaces via Ambient Compilation of Full Absorption

Christos Liaskos, George Pirialakos, Alexandros Pitilakis, Sergi Abadal, Ageliki Tsioliaridou, Anna C. Tasolamprou, Odysseas Tsilipakos, Nikolaos V. Kantartzis, Sotiris Ioannidis, Eduard Alarcón, Albert Cabellos‐Aparicio, Maria Kafesaki, Andreas Pitsillides, Kypros M. Kossifos, Julius Georgiou, Ian F. Akyildiz

Abstract

Metasurfaces constitute effective media for manipulating and transforming impinging EM waves. Related studies have explored a series of impactful MS capabilities and applications in sectors such as wireless communications, medical imaging and energy harvesting. A key-gap in the existing body of work is that the attributes of the EM waves to-be-controlled (e.g., direction, polarity, phase) are known in advance. The present work proposes a practical solution to the EM wave sensing problem using the intelligent and networked MS counterparts—the HyperSurfaces (HSFs), without requiring dedicated field sensors. An nano-network embedded within the HSF iterates over the possible MS configurations, finding the one that fully absorbs the impinging EM wave, hence maximizing the energy distribution within the HSF. Using a distributed consensus approach, the nano-network then matches the found configuration to the most probable EM wave traits, via a static lookup table that can be created during the HSF manufacturing. Realistic simulations demonstrate the potential of the proposed scheme. Moreover, we show that the proposed workflow is the first-of-its-kind embedded EM compiler, i.e., an autonomic HSF that can translate high-level EM behavior objectives to the corresponding, low-level EM actuation commands.

Introduction: The "Blind" Metasurface Problem

Metasurfaces (MS) are thin, artificial structures capable of manipulating electromagnetic (EM) waves in extraordinary ways—such as steering reflections or perfectly absorbing energy. However, most metasurfaces are "blind": they can only perform these feats if the properties of the incoming wave (like its angle of incidence or polarization) are known in advance.

Currently, sensing these properties requires bulky, expensive external sensors that lack spatial granularity. Integrating these sensors directly into the metasurface adds significant manufacturing complexity and cost. The authors address this by asking: Can we turn the metasurface itself into a sensor using its existing hardware?

The ABSense Framework: Sensing via Actuation

The researchers utilize HyperSurfaces (HSFs), which are metasurfaces embedded with a network of miniature controllers (nano-nodes). Their approach, called ABSense, treats the act of sensing as an optimization problem:

System architecture of the HyperSurface (HSF) showing the embedded nano-network and impinging wave attributes.
  1. Iterative Search: The HSF iterates through different configurations of its active elements (like PIN diodes or varactors) to change its surface impedance ($R + jX$).
  2. Maximizing Dissipation: The system looks for the specific configuration that results in Full Absorption. At this state, the power dissipated across the internal controllers is at its maximum because the surface impedance perfectly matches the incoming wave.
  3. Distributed Consensus: Because individual measurements can be noisy or erroneous, the embedded nano-nodes use a distributed consensus algorithm. They communicate with neighbors to calculate a collective average of the best configuration, making the system highly robust.
  4. Inverse Lookup: Once the "full absorption" configuration is found, the system consults a static lookup table (created during manufacturing). This table maps specific hardware states to the wave attributes that would cause full absorption in those states, effectively "reversing" the function to identify the wave.
Flowchart of the ABSense algorithm, detailing the iterative impedance search and distributed consensus process.

Key Insights and Results

ABSense represents the first workflow to act as an embedded EM compiler. It translates a high-level objective ("identify this wave") into low-level hardware commands autonomously.

  • Speed and Efficiency: The entire sensing process (measurement + consensus) takes approximately 11 microseconds. The consensus phase is highly economic, requiring only about 10 packet transmissions per neighbor to reach a stable estimate.
  • Extreme Robustness: The use of distributed consensus means the system doesn't fail if a few controllers are faulty. Simulations showed the system could still converge on the correct wave attributes even if individual node measurements had up to 82% random error.
  • Accuracy: Full-wave simulations in CST Microwave Studio and network simulations in AnyLogic demonstrated that the system accurately identifies the elevation angle ($\theta$) and distinguishes between Transverse Electric (TE) and Transverse Magnetic (TM) polarizations.

Limitations and Future Outlook

While promising, the current implementation has a few constraints. It currently works best for waves arriving in the principal planes ($xz$ and $yz$). Sensing waves from arbitrary azimuth angles ($\phi$) or with elliptical polarization would require more complex, anisotropic unit-cell designs. Additionally, the method assumes controllers can measure internal power dissipation.

Future directions include extending the system to sense non-planar, complex EM waves and developing an "always-on" sensing mode that runs in parallel with other metasurface functions like beamsteering.

Summary updated

Abstract

Metasurfaces constitute effective media for manipulating and transforming impinging EM waves. Related studies have explored a series of impactful MS capabilities and applications in sectors such as wireless communications, medical imaging and energy harvesting. A key-gap in the existing body of work is that the attributes of the EM waves to-be-controlled (e.g., direction, polarity, phase) are known in advance. The present work proposes a practical solution to the EM wave sensing problem using the intelligent and networked MS counterparts-the HyperSurfaces (HSFs), without requiring dedicated field sensors. An nano-network embedded within the HSF iterates over the possible MS configurations, finding the one that fully absorbs the impinging EM wave, hence maximizing the energy distribution within the HSF. Using a distributed consensus approach, the nano-network then matches the found configuration to the most probable EM wave traits, via a static lookup table that can be created during the HSF manufacturing. Realistic simulations demonstrate the potential of the proposed scheme. Moreover, we show that the proposed workflow is the first-of-its-kind embedded EM compiler, i.e., an autonomic HSF that can translate high-level EM behavior objectives to the corresponding, low-level EM actuation commands.
BibTeX
@inproceedings{liaskos2019absense,
  author = {Christos Liaskos and George Pirialakos and Alexandros Pitilakis and Sergi Abadal and Ageliki Tsioliaridou and Anna C. Tasolamprou and Odysseas Tsilipakos and Nikolaos V. Kantartzis and Sotiris Ioannidis and Eduard Alarcón and Albert Cabellos‐Aparicio and Maria Kafesaki and Andreas Pitsillides and Kypros M. Kossifos and Julius Georgiou and Ian F. Akyildiz},
  title = {{ABSense: Sensing Electromagnetic Waves on Metasurfaces via Ambient Compilation of Full Absorption}},
  booktitle = {NANOCOM},
  year = {2019},
  url = {https://openreview.net/forum?id=9jVq3OhOaT7},
}
Figure