A team from Queen Mary University of London and Italian universities has shown a tactile sensor for a robot fingertip that turns pressure into color. Described on 28 July 2026 in IEEE Spectrum, the sensor reaches 100-micrometer resolution with zero computational latency, because the "sensor" is the material itself, not an algorithm.
Key takeaways
- 100 µm resolution, zero computational latency.
- Principle: a mechanochromic Bragg reflector between silicone layers.
- Deformation shifts reflected-light color: red (low) to green and blue (high).
- The photosensitive film is exposed to a 635 nm laser for 7 minutes.
- Mapped a human fingertip, a U.S. penny and a leaf surface, among others.
How it works
At the core is a Bragg reflector?Bragg reflector: A stack of alternating-density layers that reflects only one wavelength of light — here shifting with the applied pressure. — a thin polymer film of alternating-density layers, placed between a protective layer and clear silicone. An LED shines through the silicone, and when contact deforms the material, the reflector bounces back light at a wavelength that depends on the pressure.
An embedded camera reads the color: red means minimal deformation, green moderate, blue maximum. The film is made by exposing a photosensitive layer to a red 635 nm laser for seven minutes, creating the alternating-density layers.
| Color | Deformation |
|---|---|
| Red | minimal |
| Green | moderate |
| Blue | maximum |
The core aspect is that we're essentially moving the sensing element to the material level.
Giacomo Sasso, postdoctoral research associate, Queen Mary University of London.
How it differs from GelSight
This isn't the first soft-material tactile sensor — the competing GelSight also builds topological surface maps. The difference is in the data. Most sensors return only a topological map showing relative feature sizes. The new device extracts quantitative depth and dimension information, mapping topology, strain and pressure at once.
Because processing happens "in the material," there is none of the compute cost typical of camera-image analysis. Rich Walker of Shadow Robot Company called it "a distinctly different approach."
Why it matters
Dexterous manipulation is still a robotics bottleneck — grasping and turning objects needs rich, fast touch. Moving part of the sensor's "intelligence" into the material lowers latency and simplifies the signal path, which matters for real-time control.
If the approach scales, it could bring robots closer to human precision in reading texture and shape, without loading the processor with heavy image analysis from every fingertip.
What's next?
- The team aims to push the sensing element to the material level — the next step is integration with a full gripper.
- Verifying the durability of mechanochromic layers under repeated, strong pressure remains an open engineering question.
- The methodology builds on mechanochromic-materials work published in Nature.
Sources
- IEEE Spectrum — Robot Finger Feels in Color
- Nature Materials — Mechanochromic material study
