Dossier 02 · Applications

Where Visuotactile Sensors Actually Get Used

Seven manipulation problems in robotics, surgery and prosthetics where visuotactile feedback has produced documented performance gains between 2020 and 2026 — with the sensors and labs behind the work.

Use Cases Profiled
7
From in-hand manipulation to surgical tool feedback — one card each.

Visuotactile sensing is rarely the answer to an entire robotic task on its own. It earns its place when a manipulation problem has a contact moment that vision alone cannot see — the instant the peg meets the hole, the millisecond the cup begins to slip, the gentle pressure threshold a surgical tool must respect. The seven cards below profile the application areas where that case has been made most clearly in published research.

Use case 01

In-Hand Manipulation

Re-orienting an object between the fingers without dropping or re-grasping it. Visuotactile feedback closes the perception loop by telling the controller which face of the object is in contact at every frame — information that an external camera typically loses to occlusion by the hand itself.

SensorDIGIT, GelSight Mini
Leading labCMU, MIT, Meta AI Research
ReferenceIn-hand cube reorientation papers, 2021-2024
MaturityLab-demonstrated, scaling
Use case 02

Slip Detection

Detecting micro-slip events at the contact patch within tens of milliseconds and triggering a grip-force adjustment before the object is lost. This is the textbook case for visuotactile sensing because the relevant signal is fundamentally a high-frequency tactile event invisible to vision.

SensorGelSight, TacTip, ReSkin
Leading labBristol Robotics Lab, MIT, NUS
ReferenceSlip-aware grasping, multiple groups 2018-2024
MaturityMature in lab, transferring to industry
Use case 03

Texture Classification

Identifying surface texture or material class from tactile contact alone. Reported accuracies on standard texture benchmarks now reach over 90 percent for visuotactile inputs, comparable to or exceeding what humans achieve in blindfolded studies.

SensorGelSight, TacTip
Leading labMIT CSAIL, Bristol, Tsinghua
ReferenceTexture-recognition benchmarks 2017-2023
MaturityStrong on benchmarks, niche in industry
Use case 04

Surgical Tool Feedback

Giving teleoperated and autonomous surgical instruments a sense of tissue contact — where today most systems rely on visual feedback alone. Visuotactile fingertips on graspers, dissectors and suturing tools allow gentle force regulation and tissue-property estimation during surgery.

SensorGelSight Wedge, miniaturised DIGIT variants
Leading labJohns Hopkins, Imperial College, Intuitive collaborators
ReferenceTactile-augmented surgical robotics, 2021-2025
MaturityPre-clinical research
Use case 05

Prosthetic Hand Feedback

Restoring a usable sense of touch to amputees through tactile sensors on prosthetic fingertips, mapped to peripheral nerve stimulation or vibrotactile cuffs on the residual limb. Visuotactile sensors have begun to replace classical capacitive arrays in research prostheses.

SensorSynTouch BioTac, GelSight variants, custom designs
Leading labUSC, Case Western, Universite Catholique de Louvain
ReferenceSensory feedback prostheses, multi-group 2020-2025
MaturityResearch, early clinical trials
Use case 06

Fragile-Object Grasping

Picking and placing thin-walled, deformable or breakable objects — egg shells, electronic components, lab glassware — without crushing them. Visuotactile feedback enables minimum-force grasping policies that adapt the grip in real time to the object stiffness.

SensorGelSight Mini, Soft-Bubble
Leading labToyota Research Institute, Stanford ILIAD
ReferenceCompliant grasping with tactile feedback, TRI 2020-2024
MaturityLab-mature, pilot deployments
Use case 07

Assembly Insertion (Peg-in-Hole)

Solving the canonical contact-rich manipulation problem: aligning and inserting a peg, plug, screw or connector with sub-millimetre tolerance. Visuotactile sensors detect edge contact, axis misalignment and friction transitions far faster than vision, often enabling insertion under tight tolerance with minimal search behaviour.

SensorGelSight Wedge, DIGIT, F/T baseline
Leading labMIT, ETH Zurich, Mitsubishi Electric Research Labs
ReferenceTactile peg-in-hole, multi-group 2019-2025
MaturityIndustrial pilots, spectrum of solutions
Maturity

Where Each Use Case Stands in 2026

The seven application areas above are not equally mature. Some — slip detection, texture classification, peg-in-hole — have a decade of cumulative evidence behind them and are entering pilot deployments. Others — surgical tool feedback, prosthetic restoration — remain primarily a research subject because they sit inside heavily regulated product domains where adoption cycles are measured in years.

The maturity bars below are an editorial estimate, not a benchmark. They are meant only to convey the relative readiness for production-grade integration as observable from public literature and industry talks at the major 2025-2026 robotics conferences (RSS, IROS, ICRA, CoRL, Humanoids).

Application
Maturity (estimated, lab → production)
Score
Slip detection
8 / 10
Texture classification
7 / 10
Peg-in-hole assembly
7.5 / 10
Fragile-object grasping
6.5 / 10
In-hand manipulation
5.5 / 10
Surgical tool feedback
3.5 / 10
Prosthetic feedback
3 / 10