cs.ROFeb 27, 2026

SpikingTac: A Miniaturized Neuromorphic Visuotactile Sensor for High-Precision Dynamic Tactile Imprint Tracking

Authors: Tianyu Jiang, Chaofan Zhang, Shaolin Zhang, Shaowei Cui, Shuo Wang

Organizations: Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China · School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing 100049, China · State Key Laboratory of Multimodal Artificial Intelligence Systems, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China · Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences, Shanghai 200031, China

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

High-speed event-driven tactile sensors are essential for achieving human-like dynamic manipulation, yet their integration is often limited by the bulkiness of standard event cameras. This paper presents SpikingTac, a miniaturized, highly integrated neuromorphic tactile sensor featuring a custom standalone event camera module, achieved with a total material cost of less than $150. We construct a global dynamic state map coupled with an unsupervised denoising network to enable precise tracking at a 1000Hz perception rate and 350Hz tracking frequency. Addressing the viscoelastic hysteresis of silicone elastomers, we propose a hysteresis-aware incremental update law with a spatial gain damping mechanism. Experimental results demonstrate exceptional zero-point stability, achieving a 100% return-to-origin success rate with a minimal mean bias of 0.8039 pixels, even under extreme torsional deformations. In dynamic tasks, SpikingTac limits the obstacle-avoidance overshoot to 6.2mm, representing a 5-fold performance improvement over conventional frame-based sensors. Furthermore, the sensor achieves sub-millimeter geometric accuracy, with Root Mean Square Error (RMSE) of 0.0952mm in localization and 0.0452~mm in radius measurement.

Figures & tables

Fig. 1: We present SpikingTac, a miniaturized neuromorphic visuotactile tactile sensor designed for seamless integration into dexterous robotic hands. By leveraging dynamic state map reconstruction for tactile imprints, SpikingTac enables high-bandwidth dynamic perception and ensures robust marker tracking under rapid physical interactions.
PerceptionSensorDimension [ mm3mm^{3} ] ↓\downarrowSensing Area [ mm2mm^{2} ] ↑\uparrowD/A RatioFPS ↑\uparrowCost []] \downarrow$
RGB-basedGelSight-Mini32×28.5×28=2553632\times 28.5\times 28=2553619×15=28519\times 15=2859025499 ∗
 GelStereo 2.030×30×29=2610030\times 30\times 29=2610023×23=52923\times 23=5294932114
Event-basedGelEvent40×40×96=15360040\times 40\times 96=15360032×32=102432\times 32=10241501802275 †
 Evetac32×33×65=6864032\times 33\times 65=6864019×15=28519\times 15=28524110005000 †
 NeuTac200×120×100=2400000200\times 120\times 100=24000002×π×102=6282\times\pi\times 10^{2}=62838203008000 †
 A. Khairi [ 31 ]116×148×88=1510784116\times 148\times 88=151078430×40=120030\times 40=12001259NA5000 †
TABLE I: Comparison of GelSight-Mini, GelStereo 2.0, GelEvent, Evetac, NeuTac, [ 31 ] and SpikingTac. ( ∗ Commodity price. † Estimated price.)
Fig. 2: Design of SpikingTac. (a) The exploded view of SpikingTac. The components labeled as 1 are screws for connecting the sensor base, the shading plate and the sensor shell. The components labeled as 6 are heat-set threaded inserts. The components labeled as 10 are 64 white circular markers arranged in an 8x8 grid. (b) The schematic diagram of SpikingTac.
Fig. 3: Comparison image of Gelsight Mini dotted gel and SpikingTac dotted gel.
Fig. 4: Overview of the proposed event-driven marker tracking pipeline. The process begins with Event Stream Capture, followed by three core phases: Geometric Reconstruction (orange), which builds and denoises the state map; Robust Calibration (grey), which aggregates multi-directional motion to determine initial reference positions; and Dynamic Tracking (green), which implements a hysteresis-aware update law with a spatial damping operator to ensure zero-baseline stability.
Fig. 5: Conceptual illustration of the event-driven marker reconstruction. Complete shape reconstruction: As the white circular marker translates by a distance equal to its diameter, the asynchronous event stream (represented by p=+1p=+1 and p=−1p=-1 polarities) cumulatively populates and clears the state map M(x,y)M(x,y) , eventually reconstructing the full geometry at the terminal position.
Fig. 6: Illustration of the systematic offsets during the marker restoration phase. Due to the viscoelastic hysteresis of the silicone elastomer, the detected centroid PdetP_{det} (red dot) fails to perfectly realign with the ideal initial reference PinitP_{init} (green dot) once the external force is removed, resulting in a consistent residual error. The red arrows indicate the rebound direction of the markers.
Fig. 7: Experimental setup for point tracking and various contact interaction modes during data acquisition.
Fig. 8: Tracking stability characterization of a single keypoint under repetitive contact tasks. (a) Ours; (b) Evetac; (c) NeuTac.
Fig. 9: Quantitative comparison of point tracking deviations across ten experimental sequences.
AlgorithmSuccess RateAverage MeanAverage Std
Evetac [ 9 ]80%1.62760.8525
NeuTac [ 11 ]100%1.42120.5742
Ours100%0.80390.4097
TABLE II: Performance Comparison of Point Tracking Algorithms.
Fig. 10: Experimental setup for collision detection. (a) GelStereo 2.0; (b) Ours.
Fig. 11: The experimental process of collision detection.
Fig. 12: Quantitative evaluation of collision detection performance under varying approach velocities. The left y-axis indicates the detection deviation ( ΔXp\Delta X_{p} ) at the instant the collision signal is triggered, while the right y-axis represents the post-impact overshoot ( ΔXe\Delta X_{e} ) after the robot reaches a complete standstill. Solid lines denote our SpikingTac sensor, and dashed lines represent the frame-based GelStereo 2.0 baseline [ 29 ] .
Fig. 13: Experimental setups for multi-scale circular geometry estimation.
Fig. 14: Schematic of the tactile cross-search strategy for hole center and radius estimation. The probe moves along the XX and YY axes sequentially to identify boundary contact points.
Fig. 15: Experimental results of multi-scale circular geometry estimation. The blue dashed lines represent the estimated circles after Kabsch alignment, while the red solid lines indicate the ground truth from the 3D model.
No.12345
Pos. Error0.09240.07560.06250.07290.0965
Rad. Error0.0159-0.02140.06630.03300.0242
No.678910
Pos. Error0.05390.10680.06710.11950.1572
Rad. Error0.0513-0.02800.1350-0.00460.1799
OverallPos. RMSE: 0.0952Rad. Mean: 0.0452
TABLE III: Quantitative results of position and radius deviation (all units in mm).
More in this topic

Tactile Sensing

Tactile sensing measures contact forces, deformation, texture, and motion to help robots perceive and act through physical interaction.

See all 138 papers

Explore similar work

CardsList
Aug 31, 2026

SpectraTac: A Compact Camera-Free Optical Tactile Sensor with Distributed Color Sensing

Tactile sensing is essential for physical interaction in robotics and human--machine systems. However, combining rich tactile information with compact hardware, low cost, and low computational overhead remains challenging. This work presents SpectraTac, a compact, camera-free optical tactile sensor that combines active red--green--blue (RGB) illumination with spatially distributed color sensing. Contact deforms a compliant transparent elastomer and modulates its internal light field, producing spatially differentiated changes in color and intensity. Three distributed color sensors capture these responses as low-dimensional spatio-spectral features, avoiding cameras, imaging optics, and high-dimensional image processing. The device measures 19.2 mm in diameter and 4 mm in height, with a material cost below USD~5. A data-driven decoding framework simultaneously estimates three-dimensional (3D) force and the contact region from the optical measurements. For 3D force prediction, the sensor achieved mean absolute errors (MAEs) of 0.161, 0.164, and 0.429 N along the x-, y-, and z-axes, respectively. The nine-region contact-classification accuracy was 99.9%. We further evaluated real-time 3D force tracking and contact-region-based human--machine interaction through an interactive control task. These results indicate that distributed color-resolved optical sensing offers a compact, low-cost alternative to camera-based tactile sensing for robotics, wearable sensing, and interactive systems.

Hao Wu, Haotian Guo, Yu Feng +3

Department of Mechanical Engineering, National University of Singapore, Singapore · Department of Mechanical Engineering, Zhejiang University, Hangzhou, China

Tactile SensingRobotic Tactile Sensing
AskPDF
Jun 2, 2026

Static and Dynamic Representations for Tactile Contact-Angle Estimation with Event-Based Sensors

Event-based tactile sensing offers low-latency signal acquisition for contact-rich robotic interaction. This paper investigates contact-angle estimation using event streams from an event-based tactile sensor (NeuroTac) and compares three event-derived spatial contour representations: a dynamic representation capturing recent event activity, a static representation recovering a more persistent contact state, and their combined representation. Across the evaluated motion scenarios, all representation pipelines exhibited P99 processing latency below 10 ms at all tested sampling intervals, demonstrating their potential for high-frequency event-based tactile angle estimation in robotic manipulation. The static representation consistently achieved marginally better performance than the dynamic and combined representations under scenario-specific training, yielding a mean overall MAE of 0.160° during continuous sensor rolling and a stop-phase mean MAE of 0.251° during randomly inserted motion interruptions. It also exhibited smaller performance fluctuations across speed and indentation depth variations than the other two representations.

Yanhui Lu, Efi Psomopoulou, Benjamin Ward-Cherrier

School of Engineering Mathematics and Technology, University of Bristol, U.K.

Tactile SensingRobotic Tactile Sensing
AskPDF
Jul 31, 2026

TacPrint: A Wearable Fingertip Tactile Sensor for Human-to-Robot Contact Reproduction

Human-centric data collection is emerging as a significant paradigm for robot skill acquisition, but seamlessly integrating low-cost, scalable tactile sensing systems that capture fine-grained fingertip interactions without compromising natural operation remains a key challenge. This reduces the reliability of human-to-robot transfer in contact-rich tasks. In this work, we present TacPrint, a wearable fingertip tactile sensor, where protrusions on the inner surface of the silicone skin are aligned one-to-one with 24 capacitive taxels to enable localized capacitive responses. A real-to-sim-to-real pipeline estimates a 35 ×\times 26 contact-depth map from 24-channel capacitive signals. Against simulation-generated labels, the model achieved a contact-region RMSE of 0.223 ±\pm 0.161 mm, a weighted-centroid error of 1.213 ±\pm 2.379 pixels, and an IoU of 0.829 ±\pm 0.169. With measured capacitive inputs, the network-predicted depth evaluated at the guide-calibrated contact center showed a mean absolute error of 0.085 ±\pm 0.057 mm across all 40 controlled trials, while the mean contact-position error was 0.250 ±\pm 0.208 mm across the 37 trials whose reference contact regions were not truncated by the sensing boundary. In human-to-robot replay, tactile-guided compensation increased grasping and wiping success rates from 0% to 91.67% and 90%, respectively. In closed-loop grasping, dense-depth feedback achieved success rates of 87.5% over all tested positions and 85% under edge-contact conditions, compared with 67.5% and 45% for raw-taxel feedback.

Yongxi Liu, Chaofan Zhang, Xingyu Zhang +4

Institute of Automation, Chinese Academy of Sciences · Imprintx Robotics · Beijing Academy of Artificial Intelligence

Robotic GraspingTactile SensingContact-Rich Robotic ManipulationTactile Sensing for Robotic Manipulation
AskPDF