Non-Imaging Optical Tactile Sensors for Robotics: Transduction Principles, System Integration, and Robotic Applications
Organizations: Department of Mechanical Engineering, National University of Singapore, Singapore · Department of Mechanical Engineering, Tsinghua University, Beijing, China · Department of Materials Science and Engineering, National University of Singapore, Singapore
Abstract
Non-imaging optical tactile sensors encode contact through changes in optical intensity, wavelength, spectra, or photon counts, offering diverse architectures for robotic touch. Their usefulness depends on whether contact variables can be reliably distinguished from coupled mechanical and environmental effects, and whether the resulting information can be delivered in time to guide action. This review develops a unified framework linking mechanical load transmission, optical encoding, readout, and tactile inference to robotic task requirements. It examines reflective and waveguide-based devices, micro/nanofibers, fiber gratings, interferometric and resonant structures, and mechanoluminescent composites. Three considerations emerge across these approaches. First, high optical sensitivity supports accurate tactile estimation only when target variables remain identifiable under the relevant contact and calibration conditions. Second, spatial coverage and feedback timing must be assessed at the system level, accounting for multiplexing, acquisition, demodulation, and inference. Third, device characterization, offline recognition, and closed-loop robotic performance provide distinct levels of evidence and require different evaluation criteria. Representative studies illustrate how these considerations shape force estimation, contact localization, slip correction and distributed interaction. This review concludes with priorities for calibration transfer, reliable interfaces, scalable readout, and task-based validation, providing a basis for selecting and developing non-imaging optical tactile architectures for robotics.
Figures & tables
| Main process | Representative device configurations | Measured signals | Tactile functions and characteristics | Refs. |
|---|---|---|---|---|
| Reflection, focusing and spot displacement | Elastic reflective elements, liquid-membrane lenses, quadrant-sensing soft pillars, integrated GaN units | Photocurrent, voltage, quadrant differential signals | Local loads and directions; force-transmission structures determine displacement, while optical paths and channel relationships determine outputs | [ 40 , 46 , 85 ] |
| Propagation loss, scattering and coupling | Soft waveguides, color-encoded elastic lightguides, optical-fiber knots, three-dimensional lightguide networks | Transmitted intensity, relative intensity, multiband outputs | Strain, curvature, contact and friction; optical-path layout and coupling relationships provide spatial information | [ 3 , 47 , 64 , 103 ] |
| Local mode fields and plasmonic resonance | Micro/nanofibers, flexible plasmonic optical fibers | Transmitted intensity | Local contact, hardness discrimination and multi-axis sensing; responses depend on fiber dimensions and flexible encapsulation | [ 101 , 65 , 107 ] |
| Bragg reflection and along-fiber backscatter | FBG arrays, weakly reflecting gratings, distributed micro/nanofibers | Reflection wavelengths, along-fiber correlation spectra | Multipoint coverage, force and location estimation; layout, strain propagation and multiplexed demodulation jointly provide spatial observations | [ 4 , 115 , 75 ] |
| Interference and resonance | MZI tactile pads, FPI diaphragms, FBG/WGM, optical-fiber rings | Spectral features or fixed-wavelength intensity | Force and location recognition, temporal texture signals, multiparameter separation; spectral features and readout methods selected according to target variables | [ 48 , 118 , 116 ] |
| Mechanoluminescence | ML composites and light-collecting optical fibers | Photon counts, temporal profiles of light peaks | Contact and slip events; direct light emission under mechanical stimulation, with outputs dependent on deformation magnitude and rate | [ 39 ] |
| Study | Metric | Value | Test conditions | Metric description |
|---|---|---|---|---|
| Transparent planar pressure-sensitive skin [ 62 ] | Normalized pressure-response slope | Maximum approximately 0.2 kPa − 1 | Inorganic source/photodiode readout; below 1 kPa; 70 mm 2 loading area | Local slope in the low-load range |
| Thin-film waveguide array [ 63 ] | Dynamic delay and hysteresis | CA 2.4 ms/6.7%; PET 2.5 ms/10.7%; PTBA 10 ms/45.1% | Identical 16 Hz, 0–3 N sinusoidal loading | Material comparison under identical conditions; delay measured in sinusoidal tests |
| Handheld stiffness probe [ 5 ] | Force and displacement RMSE | 0.08 N/0.2 mm | Tests on six materials | Continuous-variable estimation errors |
| Polymer FBG skin [ 110 ] | Force MAE and single-axis location MAE | 0.07 N; X 0.24 mm/Y 0.21 mm | Up to 3 N; random sample splitting | Prediction errors under the corresponding data split |
| Ruffini-type FBG skin [ 4 ] | Median errors on independent test data | Force 35 mN; location 3.2 mm | Complete indentation events held out | Median values, distinct from MAE |
| Helical FBG array [ 109 ] | Best force resolution reported by the authors | 0.93 mN | Best-performing element; 1 pm wavelength resolution | Derived from instrument resolution and element sensitivity |
| Technical route | Main advantages | Key limitations and system costs | Main robotic applications |
|---|---|---|---|
| Reflection, occlusion and spot displacement (LED–photodiode or quadrant units) [ 40 , 46 , 84 , 79 ] | Compact local units; directional output; simple intensity readout | Alignment-sensitive; electronics at the contact site; slower scans with many nodes | Slip correction; contact stop on robot links; proximity and tactile servoing |
| Waveguide loss and coupling (soft waveguides, lightguides, fiber knots) [ 3 , 47 , 103 , 78 , 44 ] | Soft and stretchable; simple readout; color-coded location | Bending and pressing confounded; hysteresis; device-to-device offsets; temperature effects | Softness exploration; slip feedback; soft and underwater grasping |
| Localized fields (micro/nanofibers, plasmonic fibers) [ 65 , 101 , 107 , 138 ] | High sensitivity and speed (e.g., 7 mPa, 10 µs); direction from asymmetric structures | Fragile tapers; position- and mounting-dependent response; sensitive to surroundings and bending | Hardness palpation; multi-axis and shear sensing |
| Fiber Bragg gratings with wavelength-division multiplexing [ 77 , 4 , 60 , 141 ] | Many points per fiber; remote electronics; kHz interrogation | Interrogator required; limited spectral bands; temperature cross-sensitivity | Grasp force control; contact stop on robot surfaces; instrument-tip force |
| Distributed OFDR readout (weak gratings, Rayleigh backscatter) [ 115 , 75 ] | Continuous coverage along one fiber | Tunable laser and interferometers; slow complete acquisition (about 1 s); transverse location relies on packaging | Contact localization on sensing skins |
| Interference and resonance (MZI, FPI, WGM, fiber rings) [ 119 , 116 , 120 , 48 ] | Sub-mN resolution; force–temperature separation | Temperature-sensitive spectra; spectral demodulation equipment | Instrument-tip force; texture and Braille recognition |
| Representative system and task | Task information requirements | Measured signals | Action mechanism | Demonstrated results |
|---|---|---|---|---|
| Optical-waveguide soft prosthetic hand [ 3 ] | Surface shape and softness | Optical-loss responses during scanning and pressing | Predefined open-loop exploration, followed by measurement-based object selection | Surface and softness sensing; selection of a softer tomato |
| FBG force sensor at a laparoscopic instrument tip [ 177 ] | Stiffness differences in tissue under the instrument tip | Three-axis force from four FBG wavelength shifts | Sliding along predefined paths at a fixed indentation depth; force readout without feedback to the motion | Resultant force up to 1.66 N across a simulated tumor in porcine liver; 0.40–0.63 N over normal tissue |
| Optical-fiber knot grasping [ 103 ] | Slip events during grasping | Friction-induced changes in optical intensity | Slip feedback triggers an increase in grip force | Feedback-on and feedback-off demonstrations with a progressively loaded cup |
| PapillArray grasping [ 76 ] | Incipient slip and grasp stability | Tangential force time series from multiple pillars | Arm stopping and increase to a predefined grip force upon detection | Cross-device online deployment; 36 grasp-correction trials |
| FBG robotic hand [ 77 ] | Degree of contact and changes in object dimensions | Mean grating peak shift | Threshold-based stopping of finger closure; PID control maintains the peak-shift setpoint | One breakage in 20 nacho grasping trials; tracking of dimensional changes |
| Nine-waveguide soft gripper [ 78 ] | Soft-finger bending, contact and slip | Multichannel optoelectronic voltages | Air-pump regulation through inner-channel voltage feedback; gripping parameters selected according to recognition results | Closed-loop grasping, slip adjustment and fruit-sorting demonstrations |
Tactile Sensing
Tactile sensing measures contact forces, deformation, texture, and motion to help robots perceive and act through physical interaction.
See all 154 papers