A patterned optical layer tracks three-dimensional vibrations in real time with only a few detectors

Researchers have shown that a patterned optical layer can follow a structure’s three-dimensional vibrations in real time using only a handful of detectors.

The finding suggests a lower-cost way to keep an eye on buildings and bridges for early warning signs, before conventional monitoring schemes become too cumbersome and expensive.

Light on motion

On a four-storey building model fixed to a shaker table, the set-up read movement via reflected light rather than through wired arrays of sensors.

At the University of California, Los Angeles (UCLA), Professor Aydogan Ozcan’s group demonstrated that the patterned layer could convert that motion into optical signals that can be interpreted.

Because the layer carried out the first stage of computation on its own, the full three-dimensional experiments required only four detectors.

That streamlined approach underpins the promise of the work, while also raising the obvious question: why do today’s monitoring systems still require so much equipment and computing effort?

Why current systems

Across the United States, critical public infrastructure still needs close attention, and the most recent infrastructure report awarded the country only a C.

Most structural health monitoring-ongoing checks for damage or deterioration-continues to depend on sensor networks that must be powered, serviced, and analysed.

When those networks are dense, they also generate vast volumes of readings, forcing engineers to pay to transmit, store, and process data before any problem can be identified.

That load helps explain why an optical alternative could be important, even if traditional approaches remain the default in many deployments.

Inside the patterned layer

At the centre of the device is a diffractive layer: a patterned surface designed to direct reflected light.

When a wall or beam shifts, the layer moves with it, altering the returning wave in a way that can be decoded.

In the experiments, millimetre-wave radiation-long-wavelength light between microwaves and infrared-illuminated the surface, and the motion-induced changes travelled back to the detectors.

Rather than recording unprocessed streams from numerous instruments, the method allows the light itself to perform part of the sorting, with software completing the reconstruction.

How accuracy changed

When the team evaluated different optical configurations, the jointly trained design separated vibration frequencies much more clearly than standard options.

The error in estimating vibration frequencies fell markedly when the optical layer and the decoding system were trained together.

Lens arrays and random diffusers performed worse, reinforcing that the improvement came from co-training the optics and the decoder as a single system.

That gap implies the hardware was doing more than capturing cleaner signals-it was reformulating the task into one a small network could handle.

Testing a building

For the laboratory proof-of-concept, the researchers attached a printed layer to a four-level building model.

A programmable shaker table drove the structure, while laser range finders recorded the true motion that the optical system was expected to reproduce.

The tests included one-directional and two-directional motion, spanning earthquake records, white noise, and intentional pushes applied at different storeys.

This mattered because untidy, mixed movements in the lab are closer to what real structures experience than the idealised signals often used in early engineering tests.

Power and speed

Claims about lower cost often collapse when power demand is considered, yet here the largest draw was concentrated in a single place.

Most of the energy consumption came from the light source, whereas the computational step needed only a small amount of power.

Digital reconstruction took about 72.5 million operations and completed in roughly 30 milliseconds, which is quick enough for continuous monitoring.

That division is important, because improvements in light sources could reduce overall power without requiring larger on-site computers.

Watching several points

Tracking one location is useful, but real bridges and towers require multiple points to be observed simultaneously.

To address that, the team created a wavelength-multiplexed arrangement-combining different light wavelengths-to monitor three separate points.

Once the system used enough detectors to capture every motion being tracked, accuracy increased sharply.

The outcome points towards a way to watch many locations without installing a full electronic sensor package at each measurement point.

Where it fits

Early reporting connected the approach not only to bridges and buildings, but also to disaster resilience and aerospace diagnostics.

“Our system leverages the diffractive layer as an optimized optical processor that intelligently pre-encodes complex, multi-dimensional structural oscillation information directly into modulated optical signals,” said Ozcan.

The concept is particularly relevant where cabling is difficult, available power is limited, or rapid decisions are required.

Aircraft, remote industrial facilities, and disaster zones match those constraints, although each application would need hardware tuned to its own conditions.

What still limits

Strong laboratory results do not remove the toughest challenge: building rugged versions that can operate for years outdoors on real structures.

Shifting from 3-millimetre waves to visible or infrared light would demand far finer features and tighter manufacturing control.

Any outdoor deployment would also have to withstand heat, moisture, dust, and alignment drift without disrupting the patterns the decoder relies on.

Those constraints do not weaken the underlying idea, but they set the engineering tasks that sit between a prototype and widespread public use.

Where this leads

By allowing a passive optical surface to carry out part of the sensing and processing, the work merges hardware and software into a leaner monitoring approach.

If the technology can be scaled up and made durable, continuous checks on buildings and bridges could become cheaper, faster, and much easier to deploy widely.