RoboTac 2026
IEEE/RAS International Workshop
5 June 2026
HALL C6
ICRA 2026, Vienna, Austria
The 8th
Embodied Tactile Intelligence in Predictive Perception, Learning & Control in Grasping & Manipulation
Emerging the Role of Embodiment and Visuo -Tactile - LLM Foundation Models
Call for Papers and Contributions
Important dates:
-
Paper submission deadline: April 15th, 2025
-
Notification of acceptance: April 24th, 2026
-
Camera-ready deadline: April 30th, 2026
-
Workshop day: June 5th, 2026
Paper submission guidelines:
We solicit contributions in the form of extended abstracts (min 2 pages, max 4 pages) in IEEE paper format to be presented at the workshop as posters. Please follow the authors guidelines and use the LaTex or MS Word templates for the ICRA conference. Outstanding contributions will be selected for oral presentations.
Accepted papers and eventual supplementary material will be made available on the workshop website.
The accepted papers will be invited to submit their longer version to an especial Issue Journal.
Submission:
Please submit your contribution via Email robotac@bmw.de .The subject of the email should be RoboTac2026.
In case of any questions please contact the main organiser.
Awards:
Outstanding paper, poster and demos will receive awards and certificate.
Objectives
Haptics, or the sense of touch, enables humans to perceive and interact with their environment, playing a crucial role in grasping, manipulation, learning, and decision-making. Humans rely on haptic exploration, combined with vision and higher-level reasoning, to understand an object's shape, texture, and mechanical properties, using sensory feedback to plan and adapt actions in real time.
For robots, developing similar capabilities is essential to safely and efficiently operate in dynamic and unstructured environments, from industrial tasks to household assistance and care. Achieving robust haptic exploration and object recognition remains challenging due to soft or variable materials, sensor and motor noise, and delays in sensory processing.
Embodiment is key: by grounding perception, learning, and control in the physical body and its interactions, robots can achieve more effective, interactive, and predictive manipulation. Integrating tactile sensing with vision and large language models (LLMs) allows robots to combine local, detailed touch information with global visual context and high-level reasoning. This integration enables better anticipation of object properties, more robust planning, and reasoning about complex tasks, resulting in more dexterous, adaptive, and context-aware manipulation.
Organizers
Invited Speakers and Schedule
Session 1 : Human Sense of Touch and Embodiment
09:15
Human perception of Referred Haptic feedback for Virtual Object Interactions
Abstract available soon
09:30
Decoding Human Texture Perception with Interpretable Machine Learning
When we explore surfaces through touch, our skin receives rich tactile signals that help us perceive their sensory attributes, such as smoothness, warmth, and softness. Yet, how these signals are transformed into perceptual experiences and material recognition remains poorly understood. In this talk, I will present our work using interpretable machine learning to decode tactile signals from finger-surface interactions. We employed interpretable algorithms to investigate which signal features contribute to the elicitation of perceptual attributes and the classification of materials. Furthermore, we explored whether these perceptual attributes can reliably inform material recognition. Our findings reveal decision patterns that align with human perception, offering new insights into texture recognition and supporting the development of more perception-based haptic technologies.
09:45
Soft robotics for haptic interactions
I will introduce our work in the TU/e Reshape lab on soft robotics for haptic interactions. In a constant iterative cycle of developing hardware and studying human perception, we develop series of soft robotic haptic interfaces, e.g. haptic displays based on our pneumatic unit cells (PUCs). These soft, scalable actuators can produce both static forces and vibrations, offering a pleasant and compliant alternative to traditional vibrotactile haptic feedback systems. They also serve as a versatile tool for studying human haptic perception, e.g. soft robotic thimbles for the perception of virtual button clicks, soft actuators for apparent motion studies and small soft pins for tactile displays.
10:00 Poster Teasers A
10:15 Panel Discussion Session 1 - Q/A
10:45 - Poster session 1 & Coffee Break
Session 2 : Tactile Sensing Technologies for Embodied Robots
11:00
Latest Advancements in Vision Based Tactile
Abstract available soon
11:15
A Tale of Two Touch Sensors: the Good, the Bad, and the Hard to Build
In this talk, we will introduce two touch sensors for multifingered hands developed in our lab, the PopcornFT and the SpikeATac, with a particular focus on the details of design, development and manufacturing that make a huge difference on the final performance.
11:30
Embodied Tactile Intelligence
Abstract available soon
11:45
Beyond Images and SI Units: Sensing the Task
What is the best sensing space for a given task? Presumably, it depends on the task itself. If so, why do most sensors produce images or measurements in standardized physical units? Why are sensors typically designed to maximize separability between quantities like pixels, distances, or forces, rather than directly encode task-relevant information? I will discuss sensing designed around the task. Such sensing might produce outputs that are not directly interpretable by humans, yet still provide the most informative signals for estimation or control. In this view, optimal sensing may not correspond to disentangled or semantic measurements, but rather to representations that maximize actionable information for the downstream system.
12:00 Panel Discussion Session 2 - Q/A
12:30 Launch Break & Demo Session
13:25 Reconvene for Afternoon Session
Session 3 : Embodied Visuo-Tactile & LLM Perception, Interaction & Exploration
13:55
Whole-body robot skins for interaction
I will present a synthesis of our work in these areas: (i) Use cases of large-area tactile skins; (ii) Making large area skins, (iii) Baby humanoid robots with tactile feedback, and (iv) Embodiment for manipulation is profound – transfer across touch sensors?
14:10
Meaningful touch: Perceptual invariance bridges raw tactile data to robot dexterity
Touch is a sense of mechanics. The complex deformation of the skin that occurs when manipulating objects and interacting with the world is a critical input for coordinating movement and reacting to sudden changes. The rise of high-density tactile sensors shows that this ability to sense the mechanical world is equally critical for robot manipulation. Yet interpreting raw sensor data poses a fundamental challenge: signals intertwine the influence of material properties and contact dynamics in ways that are difficult to disentangle.
Drawing on insights from human tactile perception, I show how the concept of perceptual invariance provides a principled solution. Applied to the ShadowTac, a high-density shear and normal tactile sensor, this approach allows us to estimate properties such as compliance and friction independently of object geometry, while simultaneously tracking how an object slips. With this distilled information, we demonstrate that even simple controllers are capable of maintaining stable grasps at the edge of slippage.
14:25
Magnetic-based tactile sensing for robotic manipulation.
Human dexterity remains unmatched by modern robots, yet developing more dexterous robotic systems is crucial for tackling tasks in semi-structured, unstructured, and hazardous environments.
One reason for this mismatch is the lack of a robotic sense of touch: while computer vision is a very established sensory modality for robots, tactile sensing is still underdeveloped.
For more than 20 years I have been working on robotic tactile sensing using magnetic technologies and soft materials. In the talk, I will share highlights from our recent work in tactile sensors and tactile perception, and its application to robotic haptic exploration, grasping, and manipulation, showcasing how these advancements are bringing us closer to creating truly dexterous robots.
14:40
Embodied Human Mechanical Intelligence for Robotic Tactile Sensing
Humans explore and interact with the world through the integration of computational and bodily intelligence, where each shapes the other. In touch, skin mechanics and body morphology strongly influence perception and action, providing embodied strategies to manage the complexity of tactile information generated by movement and physical interaction. These mechanisms enhance information acquisition and improve interaction effectiveness. Inspired by human sensing, this talk presents how embodied mechanical intelligence can be translated into robotics through tactile sensing technologies.
I will discuss how biologically inspired tactile sensing and interaction strategies can guide the design and control of intelligent robotic systems capable of richer, more adaptive, and more effective physical interaction with complex environments.
14:55 Poster Teasers B
15:15 Poster session 2 & Coffee Break
Session 4 : Embodied Visuo-Tactile grasp, prehensile and non-prehensile manipulation
15:40
Tactile sensing for multifingered manipulation
Dexterous manipulation is largely about establishing, sensing, and controlling wrenches through contact patches. How should tactile sensing support this process, particularly given the interacting constraints and complexity of robot hand design?
In this talk I will discuss recent efforts in the US NSF HAND Engineering Research Center on (1) benchmarking dexterity and (2) achieving in-hand dexterity with and without tactile sensing.
15:55
Interative and Predictive Tactile Robotics
Abstract available soon
16:10
Enhancing and Unifying Robotic Tactile Sensing
Generalizable algorithms for tactile sensing remain underexplored, primarily due to the diversity of sensor modalities. Recently, many methods for cross-sensor transfer between optical (vision-based) tactile sensors have been investigated, yet few works have focused on non-optical tactile sensors. In this talk, I will first discuss my lab’s work on super-resolution barometric (air-pressure-based) tactile sensing before presenting our ‘UniTav-NV’ encoder-decoder architecture to unify tactile data across non-vision-based sensors. By leveraging sensor-specific encoders, the framework creates a latent space that is sensor-agnostic, enabling cross-sensor data transfer with low errors.
16:15 Panel Discussion Session 3 & 4 - Q/A
16:45 Poster Session 3
17:15 Awards, Wrap-Up and Conclusion
Accepted Papers
-
Multifingered force-aware control for humanoid robots - P. Marra, G.M. Caddeo, U. Patticini, L. Natale
-
Grip force regulation via low-dimensional incipient slip estimation - G. Vitrani, L. Willemet, M. Wiertkewski
-
One-touch friction estimation at the onset of grasp - G. Vitrani, L. Willemet, M. Wiertkewski
-
Egocentric Tactile and Proximity Sensors as Observation Priors for Humanoid Collision Avoidance - C. Kohlbrenner, N. Pudasaini, W. Xie, N. Sivagnanadasan, N. Correll, A. Roncone
-
On the Role of Haptic Feedback in Demonstration Collection for Visuomotor Policy Learning - J. Peng, A.Schreder, Y. Liu, M. Dola, B. Tran, Y. Hu, P. Xi
-
TactileHD: A Resource-Efficient HDC-based Framework for Multitask Tactile Perception and On-Device Active Learning - J. Zhang
-
TouchDrive: Electronics-Free Tactile Sensing Interface for Assistive Grasping - J. Xu
-
Variable Stiffness Tactile Sensor with Magnetically Reprogrammable Material - C. B. Akkaya, C. P. Avdar, E. Samur
-
Core-Agnostic Compliance Perception for Rigid–Deformable Coupled Objects using Vision-Based Tactile Sensing - C. Zhao, Y. Ding, H. Zhao, Y. Hu, D. Ma
-
Tactile Perception through Fluid-Solid Interaction - A. Goshtabi, M. Berghuis, A. Parvaresh, S. P. M. Babu, R. W. Srtyle, A. Rafsanjani
-
Active Tactile Pose and Shape Estimation of Highly Dynamic Objects - E. K . Gordon, B. Baraki, H. Bui, M. Posa
-
FARM: Force-Aware Robotic Manipulation with Tactile-Conditioned Diffusion Policies - E. Helmut, N. Funk, T. Schneider, C. de Farias, J. Peters
-
Haptic Affordance Learning and Exploration of Ambiguous Objects - R. E Kaundinya, B. Maacaron, S. F. Muller-Cleve, A. Glover, J. Borst, C. Bartolozzi, N. Taatgen
-
Characterisation of a Monolithic 3D-printed Tactile Sensor Using an SSIM-based Analysis - X. Guo
-
Contrastive Alignment of Force/Torque Time Series and Language using Temporal Transformers - K. Takahashi, M. Ishige, T. Tsumura, N. Megahan, T. Suzuki, A. Adziman, R. Sakuma, H. Murakami, M. Kamezaki, Y. Kawahara
-
Capturing Delicate Tactile Interactions with High Spatio-Temporal Resolution - T. Daunizeau, B. Pasquale, H. Shea.
-
Electroadhesive Gripper with Thermochromic Visuo-Tactile Sensing - T. Daunizeau, H. Shea
-
Dense Force Estimation with an Event-based Optical Tactile Sensor - A. Politis, R. Zurbrügg, V. Cavinato
-
PofTac: A Robust Plastic Optical Fiber-based Tactile Sensor for Reliable Human-to-Robot Skill Transfer - C. Chung, K. Kyung
Topics of Interest
(but not limited to)
Human Sense of Touch and Embodiment
• Touch physiology: from skin to brain, and its role in embodied intelligence
• Haptic perception and sensory-motor integration
• Action-perception loops for adaptive behavior
• Perception for learning in an embodied context
• Insights from human embodiment for robotic interactive and predictive perception
Tactile Sensing Technologies for Embodied Robots
• Conformable, compliant, and biomimetic materials
• Features enabled by embodied tactile sensors
• Sensor effects, self-healing properties, and strategies for robust robotic embodiment
• Sensor skins: design, fabrication, and integration for embodied manipulation
• Integration and read-out strategies in physically embodied robots
• Enabling technologies for fully integrated, multimodal robotic systems
Embodied Visuo-Tactile & Language: Perception, Interaction, and Exploration
• Exploitation of contact constraints and novel contact models in embodied contexts
• Object perception and exploration through touch and vision
• Tactile information processing and feature learning for embodied reasoning
• Tactile-based object modeling, localization, shape reconstruction, and classification
• Integration of vision and touch sensing with foundation models (LLMs, multimodal models) for reasoning, prediction, and planning
• Modeling and representation of embodied, multimodal sensory modalities
Embodied Visuo-Tactile & Language: Grasping and Manipulation
• Slip detection, grasp planning, and stability assessment in embodied systems
• Soft, in-hand, and whole-body manipulation
• Tactile-vision integration for predictive planning and control
• Embodied skill transfer, tactile transfer learning, and learning from human demonstrations
• Multi-robot manipulation, coordination, and collaborative assembly with embodied reasoning
• Whole-body, multi-contact planning and control
• Design and characterization of contact-exploiting, compliant robotic hands
• Leveraging foundation models to enhance dexterous manipulation, high-level task reasoning, and adaptive control
Sponsors
Gold Partners
Silver Partners
Bronze Partners
Partners
