Research in progress
RoboSkin Working Papers
Explore research proposals developed for RoboSkin.ai, with versioned manuscripts, reproducible experiments, and clear evidence status.
Browse working papers ->Research
Tactile robotics research: methods, results, and limitations
Read paper reviews across robot skin, tactile sensing, manipulation, and robot learning. Each review explains the method, what the authors measured, where the evidence stops, and which primary sources to consult before attempting a reproduction.
Public notes use conservative language. Application-specific performance, durability, and integration claims still need source-backed, application-specific evidence.
- 44
- technical briefs
- 23
- research lanes
- 44
- source paths
- 165
- focus terms
↳About the research, sources & related routes
Source boundary
Public sources stay visible
Each note keeps source attribution separate from RoboSkin.ai interpretation.
Evaluation lens
Claims are routed to checks
Performance, durability, and integration language is framed as evidence to verify.
Organization map
Models connect to verified institutions
Trace source-listed universities, research labs, and companies to the robot AI models and primary evidence that support each relationship.
Browse research organizations ->Physical AI route
Tactile research supports physical-world AI
Use the Physical AI explainer to connect robot skin research, tactile sensing, and contact feedback without turning source notes into product claims.
Open Physical AI contact-feedback route ->Touch data route
Contact logs make research reusable
Use the touch-data guide to connect tactile papers to contact events, timestamps, body frames, calibrated values, and robot actions.
Trace Physical AI touch data ->Project history
Separate the EU ROBOSKIN project from RoboSkin.ai
Use the source-backed FP7 project record for grant dates, funding, partners, the 2013 Springer chapter, and a clear no-affiliation boundary.
Open the EU ROBOSKIN project record ->↳Browse 23 research topics
Research lane
Conformable Skin
2026-06-17 / 5 min read
DexSkin and high-coverage conformable robotic skin for manipulation
A practical research note on DexSkin, conformable capacitive e-skin, high-coverage tactile sensing, calibration transfer, and contact-rich manipulation learning.
DexSkinconformable robot skincontact-rich manipulationcalibration transfer
Research lane
Dexterous Manipulation
2026-04-23 / 5 min read
Full-hand tactile sensing for adaptive dexterous grasping
A research-oriented explanation of why full-hand tactile coverage matters for adaptive grasping, occluded manipulation, and contact-aware robot hands.
full-hand tactile sensingadaptive graspingdexterous manipulationhigh-resolution touch
Research lane
Field Map
2026-06-18 / 5 min read
Tactile Robotics outlook for robot skin research priorities
A source-backed research landscape note on Tactile Robotics outlook, sensor types, distributed tactile sensing, simulation, benchmarking, and data interpretation.
Tactile Robotics outlookdistributed tactile sensingtactile benchmarkingrobot skin research map
Research lane
Human-Robot Interaction
2026-06-16 / 5 min read
Fluid-based robot skin for contact detection and thermal display
A source-backed note on fluid-based soft robot skin, human-robot interaction, contact detection, temperature control, and why exterior feel matters.
fluid-based robot skinthermal displaycontact detectionhuman-robot interaction
Research lane
Humanoid tactile sensing
2026-08-18 / 8 min read
Tac4Loco uses plantar pressure to adapt humanoid locomotion
Tac4Loco turns bilateral plantar pressure maps into post-contact feedback for Unitree G1 locomotion on slopes, partial support, foam, and gravel.
Tac4Locoplantar pressure sensinghumanoid locomotionUnitree G1
Research lane
Large-Area E-Skin
2026-06-18 / 5 min read
Origami capacitive robotic e-skin for large-area tactile sensing
A source-backed brief on origami capacitive robotic e-skin, large-area coverage, super-resolution tactile localization, shear sensing, and proximity detection.
origami capacitive robotic e-skinlarge-area tactile sensingsuper-resolution localizationproximity sensing
Research lane
Magnetic Tactile Sensing
2026-06-18 / 5 min read
Open-source magnetic tactile calibration for gripper-agnostic touch
A practical research note on open-source magnetic tactile calibration, three-axis force sensing, in-situ calibration, and why low-cost sensors still need repeatable setup.
open-source magnetic tactile calibrationthree-axis force sensingin-situ calibrationrobot grippers
2026-06-18 / 5 min read
eFlesh and customizable magnetic touch sensing for robot grippers
A source-backed note on eFlesh, cut-cell microstructures, 3D-printable magnetic tactile sensors, low-cost fabrication, and robot gripper deployment tradeoffs.
eFleshmagnetic tactile sensing3D-printable sensorsrobot grippers
Research lane
Manufacturing
2026-04-18 / 5 min read
Large-area flexible tactile arrays for curved robot surfaces
A deployment-focused article on large-area tactile arrays, curved-surface coverage, adjustable resolution, slip detection, and manufacturing tradeoffs.
large-area tactile arraycurved robot surfacesslip detectionmanufacturing tradeoffs
Research lane
Multimodal E-Skin
2026-06-18 / 5 min read
Optical/electronic artificial skin for molecular sensing
A research note on optical/electronic artificial skin, CNT haptic layers, near-infrared molecular sensing, force-temperature sensing, and robotic perception beyond touch.
optical/electronic artificial skinmolecular sensingnear-infrared sensingmultimodal e-skin
Research lane
Multimodal Sensing
2026-04-22 / 5 min read
Temperature/pressure bimodal sensing and the crosstalk problem
A practical brief on temperature and pressure bimodal tactile sensing, signal decoupling, and why crosstalk matters for robot skin.
temperature/pressure bimodalsignal decouplingcrosstalkmultimodal e-skin
Research lane
Neuromorphic Touch
2026-06-18 / 5 min read
Energy constrained touch encoding for large-area e-skin
A technical brief on bioinspired spiking architecture, large-area soft e-skin, low-power tactile localization, and neuromorphic touch processing.
energy constrained touch encodingspiking neural networklarge-area e-skinneuromorphic tactile sensing
Research lane
Robot learning
2026-09-19 / 7 min read
SlipSense: from pressure and vibration to a timely regrasp
Pressure and vibration improve slip recognition together. Independent motion labels and five failed recoveries explain why detector quality and controller quality need separate tests.
slip detectionpressure and vibrationdetection latencyregrasp control
2026-09-19 / 8 min read
Touch2Trace: what makes a tactile cable-tracing policy work
A 60 Hz tactile policy traces farther than a joint-only baseline, but the useful lesson is how sensor detail, pretraining, temporal context and action timing interact.
tactile policycable tracingtemporal contextencoder pretraining
2026-09-19 / 8 min read
Visible Touch: how contact overlays reach a visual robot policy
Rendering contact as image arrows reuses a visual policy’s input pathway. The hardware and ablations show why spatial alignment, signal normalization and training still matter.
contact overlaysvisuomotor policymagnetic tactile sensingsensor-camera calibration
2026-09-18 / 7 min read
DexTouch-WM learns robot contact dynamics from human touch
Human tactile pretraining improves the paper’s contact predictions, but synthetic training data does not consistently improve real-robot policy scores. Here is what each experiment measures.
tactile world modelshuman-to-robot demonstrationscontact predictionsynthetic robot data
2026-09-18 / 8 min read
Bench2Dex benchmarks bimanual robots with simulated touch
Bench2Dex connects teleoperation, replay, tactile maps, and evaluation across 12 hand embodiments. We examine the data contract and distinguish simulation access from physical sensor validation.
tactile simulationbimanual manipulationrobot dataset validationbenchmark protocols
2026-08-22 / 8 min read
ADEPT reports a 3/10 to 8/10 tactile ablation on dexterous insertion
ADEPT reports 3/10 vision-only versus 8/10 visuo-tactile final success in one matched Flexiv-Sharpa insertion condition, with ten physical trials per condition.
ADEPTreinforcement learningdexterous manipulationvisuo-tactile policy
Research lane
Robot teleoperation
2026-08-22 / 8 min read
The Missing Touch tests spatial tactile feedback in robot teleoperation
A GelSight Mini and 32-DoF fingertip display made two teleoperation tasks more natural and consistent, but no autonomous robot policy was trained or evaluated.
robot teleoperationspatial tactile feedbacktactile sensingGelSight Mini
Research lane
Robotics Software
2026-04-19 / 5 min read
ROS 2 tactile sensor pipeline for robot skin data replay
A robotics software guide to ROS 2 tactile data messages, rosbag workflows, force-torque context, and replayable robot skin evaluation.
ROS 2 Kiltedros2_controlforce/torque broadcastersensor data pipeline
Research lane
Slip Detection
2026-06-18 / 5 min read
Wet slippage detection for bionic fingertip e-skin
A source-backed note on AI-integrated bionic fingertip e-skin, wet slippage detection, fingerprint microtextures, and why dry-surface slip claims are not enough.
wet slippage detectionbionic fingertip e-skinfingerprint microtexturerobotic hands
2026-06-18 / 5 min read
Slip-actuated bionic tactile sensing with E-textile
A source-backed note on slip-actuated bionic tactile sensing, dynamic DC generator E-textile, robotic fingers, and real-time grasp monitoring.
slip-actuated bionic tactile sensingE-textiledynamic slip detectionrobotic fingers
Research lane
Soft E-Skin
2026-04-24 / 7 min read
Single-material soft robotic skin for multimodal e-skin sensing
A peer-reviewed single-material robotic skin uses wrist-mounted EIT electrodes and data-driven channel selection to interpret touch, strain, heat, damage, environment, and proprioception across a soft hand.
single-material soft robotic skinelectrical impedance tomographymultimodal touchproprioception
2026-04-20 / 5 min read
Self-healing multimodal e-skin: useful direction, careful claims
A conservative guide to self-healing e-skin claims, multimodal sensing, damage recovery, and what must be validated before deployment language is credible.
self-healing e-skinmultimodal sensingdamage recoverysoft materials
Research lane
Tactile AI
2026-08-22 / 17 min read
UniVTAC separates tactile simulation, representation learning, and policy evaluation
Choose the matching code branch and dataset before using UniVTAC. This guide separates its tactile encoder, simulation benchmark, public downloads and paper-reported physical results.
UniVTACvisuo-tactile simulationtactile representation learningtactile robotics benchmark
2026-08-22 / 10 min read
Vision-based tactile intelligence connects sensor optics to robot action
A 2026 review maps vision-based tactile sensing as one integrated stack: deformable contact hardware, optical readout, tactile representations, learning, simulation, datasets, and robot action.
vision-based tactile sensorstactile AIoptical tactile sensingtactile foundation models
2026-08-15 / 6 min read
FeelWorld predicts contact, tactile force states, and slip for robot planning
FeelWorld adds explicit contact, force-related tactile, and slip prediction to a visual world model for contact-rich robot planning.
FeelWorldvisuo-tactile world modelcontact-rich manipulationtactile world model
2026-08-05 / 8 min read
HT-Bench full-hand tactile benchmark for robot manipulation
HT-Bench v2 pairs egocentric vision with millions of full-hand tactile frames, corrects the vision-to-tactile metric split, and adds four real-robot evaluations.
HT-Benchfull-hand tactile sensingegocentric visiontactile representation learning
2026-06-18 / 5 min read
Sparsh-X multisensory touch representations for tactile AI
Sparsh-X fuses image, audio, motion, and pressure from Digit 360, showing how multisensory touch can improve tactile AI for robot manipulation.
Sparsh-Xmultisensory touchDigit 360self-supervised tactile representation
2026-06-18 / 5 min read
MiTaS multi-resolution tactile imitation learning for robot hands
A technical note on MiTaS, heterogeneous tactile sensors, GelSight and event-based touch fusion, and why tactile frequency matters for robot hand learning.
MiTaSmulti-resolution tactile sensingevent-based touchGelSight
2026-06-18 / 5 min read
Dream-Tac: A Unified Tactile World Action Model for Contact-Rich Robot Manipulation
Dream-Tac models action-conditioned tactile futures for contact-rich robot manipulation, showing why robot skin data needs prediction, not only reaction.
Dream-Tactactile world modelcontact-rich manipulationpredictive tactile control
2026-06-18 / 5 min read
GenForce transferable force sensing for robot skin and tactile sensors
GenForce explores transferable force sensing across tactile sensors, reducing repeated calibration work for robot skin replacements and hardware changes.
transferable force sensingGenForcecross-sensor calibrationslip detection
2026-04-21 / 5 min read
Event-based tactile sensing for sparse, low-latency robot touch
A source-backed note on event-based and neuromorphic tactile sensing, sparse contact events, and why low-latency touch matters for robot control.
event-basedneuromorphic tactile sensingopto-tactile skinlow-latency touch
Research lane
Tactile benchmarks
2026-08-22 / 8 min read
SoftVTBench separates deformable-task completion from contact quality
SoftVTBench pairs simulated visual and tactile observations with evaluator-only FEM state; its Hugging Face card and GitHub README currently disagree about the hosted release scale.
SoftVTBenchvisuo-tactile datasetdeformable-object manipulationtactile benchmark
2026-08-21 / 7 min read
TactiDex benchmarks contact-level human-to-robot dexterity
TactiDex aligns whole-hand tactile signals with kinematic and object states to evaluate physically grounded human-to-robot dexterous transfer.
TactiDexTactiSkilltactile benchmarkhuman-to-robot transfer
Research lane
Tactile Data
2026-06-18 / 5 min read
FreeTacMan robot-free visuo-tactile data collection for tactile AI
A research note on FreeTacMan, robot-free visuo-tactile datasets, tactile AI data collection, and why robot skin models need contact diversity.
FreeTacManvisuo-tactile data collectionhuman demonstrationscontact-rich manipulation
2026-06-18 / 8 min read
GIST humanoid visual-tactile-action dataset maps 101.9K soft-object samples
The GIST preprint reports 101.9K visual-tactile-action samples for towel and sponge manipulation, with dense hand touch, two camera views, and explicit access limits.
GISThumanoid visual-tactile-action datasetInspire RH56-DFXsoft-object manipulation
Research lane
Tactile datasets
2026-09-11 / 8 min read
TacVerse tests whether touch transfers between sensors
TacVerse compares shape, grating, and force perception across tactile sensors. Its public code and gated dataset are now identifiable, but transfer quality depends strongly on the task.
cross-sensor tactile transfertactile datasetsvision-based tactile sensingbenchmark splits
2026-08-22 / 8 min read
PRISM maps 5,000+ contact-rich industrial robot trajectories
PRISM reports 5,000+ robot trajectories across 25+ industrial tasks, but tactile observations cover only a subset and the announced dataset download is not yet available.
PRISM datasetrobotics datasetcontact-rich manipulationrobot teleoperation
2026-08-21 / 9 min read
RoboTacDex maps 6,000+ humanoid visual-tactile trajectories
RoboTacDex reports 6,000+ Unitree G1 trajectories across 19 tasks, an author-reported 23 skills, and 22 objects, but public dataset access remains pending.
RoboTacDexUnitree G1humanoid datasetvisual-tactile-action data
Research lane
Tactile Hardware
2026-04-25 / 5 min read
Graphene and liquid metal 3D force sensing for robot fingertips
A source-backed technical brief on miniaturized tactile sensors that separate normal force, shear force, slip, and texture signals for dexterous robot hands.
graphene and liquid metal3D force sensingslip detectionrobot fingertips
Research lane
Tactile manipulation
2026-09-11 / 7 min read
TacPrint reconstructs contact from 24 tactile cells
A wearable capacitive fingertip sensor predicts a dense contact map for human-to-robot replay. We separate simulation labels, physical measurements, and closed-loop grasping results.
capacitive tactile sensingcontact reconstructionrobot handshuman-to-robot demonstrations
2026-08-21 / 8 min read
T-Rex adds high-rate tactile reaction to dexterous robot policies
T-Rex reports a 100-hour collection and a 30-point gap over EgoScale; its official release now provides code, checkpoints, and an approximately 50-hour public subset.
T-Rextactile-reactive manipulationdexterous manipulationVLA
Research lane
Tactile world models
2026-08-21 / 7 min read
HiTac-WAM forecasts contact, deformation, and slip before robot action
HiTac-WAM ranks candidate robot actions with hierarchical tactile forecasts, then checks predicted touch against measured touch during execution.
HiTac-WAMtactile world action modelcontact predictionslip prediction
Source correction path
Need more context?
Share a source, correction, or terminology route so the site can stay accurate and clearly framed.