Haptic feedback can convey spatial information without engaging vision or hearing, making it well suited for eyes-free and assistive applications. However, many haptic guidance systems rely on cues that require explicit cognitive interpretation, which can slow user responses and introduce ambiguity in the conveyed spatial information. To address this, we introduce fingertip deflection as a kinesthetic guidance cue via the NURing, a wearable tendon-driven device that delivers controlled fingertip deflections to guide whole-arm motion, designed to leave the fingerpad unobstructed for tactile exploration. Inspired by the instinctive response of toddlers guiding parents by the finger [1], deflection cues provide a mechanically direct and readily interpreted form of haptic guidance. Two user studies evaluated closed-loop guidance in horizontal (single-axis, Study 1) and horizontal-vertical (dual-axis, Study 2) task spaces. In singleaxis tasks, blindfolded participants reached a median endpoint error of 11.3mm in under 3 s on average, and tracked moving targets with 53.5mm mean error. In dual-axis tasks, participants accurately identified deflection cue directions, and successfully reached two-dimensional target poses with a median endpoint error of 12.6mm in under 7 s on average. Kinematic analysis revealed that participants responded rapidly (median 330 ms), with early trajectories biased within 25° toward the target heading, indicating that deflection cues provided early, actionable direction information. These results demonstrate that fingertip deflection via the NURing enables fast, readily interpreted kinesthetic guidance that produces a coordinated whole-arm response without obstructing tactile contact, supporting its potential for eyes-free interaction and assistive applications.

Tactile sensing plays a key role in enabling dexterous and reliable robotic manipulation, but realizing this capability requires substantial calibration to convert raw sensor readings into physically meaningful quantities. Despite its near-universal necessity, the calibration process remains ad hoc and labor-intensive. Here, we introduce 3D Cal, an open-source library that transforms a low-cost 3D printer into an automated probing device capable of generating large volumes of labeled training data for calibrating vision-based tactile sensors. 3D Cal also provides an end-to-end, user-friendly pipeline for training custom convolutional networks to produce high-quality depth reconstructions. Using 3D Cal, we systematically explore the relationship between training data volume and spatial reconstruction performance on two commercially available sensors, DIGIT and GelSight Mini, and derive practical, empirically-grounded guidelines for calibrating these sensors. Finally, we demonstrate depth reconstruction performance on the DIGIT and GelSight Mini comparable to state-of-the-art methods, achieving average reconstruction errors of 156 μm and 205 μm on unseen objects, respectively. By automating tactile sensor calibration, 3D Cal can accelerate tactile sensing research, simplify sensor deployment, and facilitate the integration of tactile sensing in robotic platforms.

Many eyes-free guidance systems convey directional information through symbolic cues requiring interpretation, potentially increasing cognitive load and competing with existing sensory channels. We present an interactive demonstration of fingertip deflection as a continuous, physically-intuitive guidance cue that gently biases the arm during reach. Motivated by a challenge frequently reported by individuals with blindness — that of safely retrieving small objects within arm’s reach — we developed an eyes-free, object-retrieval task. Building on our prior NURing, we extend fingertip deflection beyond instrumented testbeds by integrating a camera on the actuation ring. This allows the NURing to detect ArUco-tagged objects, estimate their pose relative to the camera, and drive continuous closed-loop deflection cues toward the target in real time. This demonstration invites attendees to experience this embodied guidance firsthand and to explore how fingertip deflection could support future assistive and collaborative systems that guide action through physical intuition rather than cognitive translation.

Dexterous manipulation in compact robots requires combining high force output with passive backdrivability; capabilities that conventional geared actuation struggles to deliver. We introduce an electromechanical multiplexing architecture that routes power from a single drive shaft to multiple outputs and mechanically grounds them using capstan-amplified electroadhesive (EA) clutches in a load-transfer configuration. Wrapping thin-film EA clutches on cylindrical counter-surfaces provides exponential gain for EA braking force, while voltage pulse-width modulation yields sub-newton (<0.1 N) force resolution and low reflected impedance from the drive shaft, enabling compliant interaction. Force transmission behavior is explained by a mechanics-based model of curved clutches and supported by strain imaging showing a propagating, load-carrying slip front. Switching measurements under bipolar high-voltage drive demonstrate millisecond-scale release and effective operation near 1 kHz, enabling high-rate force modulation. Leveraging this clutch-level understanding, we realized a well-behaved system: a tendon-driven, two-finger gripper that transitions between highly backdrivable, cooperative grasping and firm, energy-efficient holding. By decoupling pulling from latching, the load-transfer design mechanically grounds the output without sustained motor torque, outlining a scalable route to compact, low-power robotic hands that maintain backdrivability while spanning three orders of magnitude in force.

Robot hands are a key interface between AI and the physical world, making advances in robotic dexterity essential to realizing the vision of physical AI. While impressive dexterity has been demonstrated with simple grippers, multifingered hands offer the potential for substantially greater versatility, precision, and adaptability in manipulation. In this review, we survey the state of the art in benchmarking the dexterity of multifingered robot hands. Recognizing dexterity as a complex and multifaceted concept, we present the perspective of the U.S. National Science Foundation HAND Engineering Research Center, with a particular focus on fine in-hand manipulation. We introduce a framework consisting of three benchmark levels that correspond to increasing system complexity, review representative benchmarks at each level, and propose new benchmarks and metrics to address limitations in the literature. More information can be found at https://hand-erc.github.io/benchmarking/.

Many eyes-free guidance systems convey direction through symbolic cues that require interpretation, potentially increasing cognitive load and competing with critical sensory channels. This work builds on prior studies of fingertip deflection as a physically-grounded guidance cue previously demonstrated in controlled 1D and 2D settings using an instrumented testbed. Here, a camera is integrated onto the actuation ring to enable object-referenced, closed-loop guidance during free reach. Using ArUco targets, the system estimates target displacement in the camera frame and maps this error into continuous fingertip deflection cues in real time. Two pilot demonstrations illustrate feasibility: (1) a Fitts-style touchscreen target acquisition task with trajectory visualization, and (2) an object-retrieval task with video-based evidence of guided approach and grasp. These vignettes ground a forthcoming study with participants who are blind and vision-impaired, and invite discussion on how ring-mounted sensing can best support embodied, eyes-free guidance in everyday interactions.

Autonomous robot policies are commonly trained using demonstration data acquired via robotic teleoperation, a process which can be time-intensive and physically demanding for human operators. Bilateral control can speed up robotic teleoperation by allowing the operator to feel the forces experienced by the remote manipulator, but it can also transmit undesirable forces, such as friction and damping, back to the operator. Here, we present dead zone bilateral control, an improvement to the conventional position-position bilateral control scheme that prevents the transmission of unwanted forces during free-space motion. We implement our controller on a custom 2-degree-of-freedom teleoperation device and show that it reduces the energy required for free-space motion by approximately 50%. During a user study in which 16 participants performed a peg rolling task, our controller reduced completion times by an average of 25% compared to the standard position-position bilateral controller. The results suggest that dead zone bilateral control can expedite the collection of teleoperated task demonstrations, allowing researchers to gather larger datasets for training autonomous robot policies.

A fundamental challenge in robotic teleoperation is enabling an operator to control a remote robot as effortlessly and intuitively as their own hands. Despite the growing use of teleoperation to collect demonstration data for training autonomous robot policies, teleoperated robot performance still falls significantly short of human dexterity, even for basic tasks. Here, we present evidence that a key factor contributing to this performance gap is the absence of spatially distributed tactile feedback. Using a two-degree-of-freedom (DoF) bilateral force-feedback telemanipulator paired with a 32-DoF tactile fingertip display, we show that operator performance improves significantly when localized deformations on the remote manipulator are faithfully reproduced on the operator's fingertip. In a series of teleoperation tasks, reproducing distributed contact information not only accelerated task performance but also brought teleoperated movements closer to natural human behavior by minimizing corrective actions and task completion steps, thereby reducing the deviation between teleoperated and natural trajectories by 29x201379

Despite advances in digitizing vision and hearing, touch still lacks an equivalent digital interface matching the fidelity of human perception. This gap limits the quality of digital tactile information and the realism of virtual experiences. Here, we introduce a step toward human-resolution haptics: a class of wearable tactile displays designed to match the spatial and temporal acuity of the human fingertip. Our device, VoxeLite, is a 0.1-millimeter-thick, 0.19-gram, skin-conformal array of individually addressable soft electroadhesive actuators ("nodes"). As users touch and move across surfaces, VoxeLite delivers high-resolution distributed forces via the nodes. Enabled by scalable microfabrication techniques, the display achieves actuator densities up to 110 nodes per square centimeter, produces stimuli up to 800 hertz, and remains transparent to real-world tactile input. We demonstrate its ability to render small-scale hapticons and virtual textures and transmit physical surfaces, validated through human psychophysics and biomimetic sensing. These findings position VoxeLite as a platform for human-resolution haptics in immersive interfaces, robotics, and digital touch communication.

In this article, we characterize the passivity of a class of haptic systems modeled as a simple sampled-data system. We guarantee passivity by ensuring that there is sufficient damping in the haptic interface. Previous work established a necessary and sufficient bound on damping, but the corresponding mathematical expressions were complicated, and the derivation was not completely rigorous. After providing a rigorous proof, we derive a more tractable expression. Using this improved expression, we establish passivity conditions for several classes of transfer functions representing virtual environments, including some special cases with time delay. The original results assumed that the operator can be modeled by a passive but otherwise arbitrary transfer function. This assumption is weakened to allow the operator model to have a shortage of passivity. This requires only a slight modification of the passivity bound.

Electroadhesive (EA) clutches are promising for advanced motion and force control in robotics, haptics, and rehabilitation, owing to their compactness and light weight. However, their practical use is limited by the inability to deliver high forces at low voltages, primarily due to a lack of understanding of their mechanics. We introduce a novel deformable body fracture mechanics approach and high-resolution strain field imaging to reveal that nonuniform stress distributions cause EA clutches to fail through delamination and crack propagation. Using this insight, we developed EA clutches sustaining 22 newtons over 1 square centimeter at 100 volts, achieving the highest stress per voltage among similar clutches. This was achieved by incorporating a soft interlayer and peeling stopper for uniform stress distribution and mitigating the failure modes. These EA clutches were integrated into a lightweight ring-based wearable system for finger rehabilitation and haptics. Our findings lay the groundwork for designing low-voltage, high-performance EA clutches for next-generation motion and force control applications.

Generating salient and intuitively understood haptic feedback on the human finger through a non-intrusive wearable remains a challenge in haptic device development. Most existing solutions either restrict the hand and finger's natural range of motion or impede sensory perception, quickly becoming intrusive during dexterous manipulation tasks. Here, we introduce NURing (Non-intrUsive Ring), a tendon-actuated haptic device that provides kinesthetic feedback by deflecting the finger. The NURing is easily donned and doffed, enabling on-demand kinesthetic feedback while leaving the hand and fingers free for dexterous tasks. We demonstrate that the device delivers perceptually salient feedback and evaluate its performance through a series of uniaxial motion guidance tasks. The lightweight NURing device, measuring approximately 220 g, can generate guidance cues at up to 1 Hz, enabling participants to identify target directions in under 3 s with a 1.5 degrees steady-state error, corresponding to a fingertip deviation of less than 11mm. Additionally, it can guide users along complex, smooth trajectories with an average trajectory error of 7 degrees. These findings highlight the effectiveness of fingertip deflection as a kinesthetic feedback modality, enabling precise guidance for real-world applications such as sightless touchscreen navigation, assistive technology, and both industrial and consumer augmented/virtual reality systems.

The ability to render realistic texture perception using haptic devices has been consistently challenging. A key component of texture perception is roughness. When we touch surfaces, mechanoreceptors present under the skin are activated and the information is processed by the nervous system, enabling perception of roughness/smoothness. Several distributed haptic devices capable of producing localized skin stretch have been developed with the aim of rendering realistic roughness perception; however, current state-of-the-art devices rely on device fabrication and psychophysical experimentation to determine whether a device configuration will perform as desired. Predictive models can elucidate physical mechanisms, providing insight and a more effective design iteration process. Since existing models (1, 2) are derived from responses to normal stimuli only, they cannot predict the performance of laterally actuated devices which rely on frictional shear forces to produce localized skin stretch. They are also unable to predict the augmentation of roughness perception when the actuators are spatially dispersed across the contact patch or actuated with a relative phase difference (3). In this study, we have developed a model that can predict the perceived roughness for arbitrary external stimuli and validated it against psychophysical experimental results from different haptic devices reported in the literature. The model elucidates two key mechanisms: (i) the variation in the change of strain across the contact patch can predict roughness perception with strong correlation and (ii) the inclusion of lateral shear forces is essential to correctly predict roughness perception. Using the model can accelerate device optimization by obviating the reliance on trial-and-error approaches.

The rich set of mechanoreceptors found in human skin1,2 offers a versatile engineering interface for transmitting information and eliciting perceptions3,4, potentially serving a broad range of applications in patient care5 and other important industries6,7. Targeted multisensory engagement of these afferent units, however, faces persistent challenges, especially for wearable, programmable systems that need to operate adaptively across the body8-11. Here we present a miniaturized electromechanical structure that, when combined with skin as an elastic, energy-storing element, supports bistable, self-sensing modes of deformation. Targeting specific classes of mechanoreceptors as the basis for distinct, programmed sensory responses, this haptic unit can deliver both dynamic and static stimuli, directed as either normal or shear forces. Systematic experimental and theoretical studies establish foundational principles and practical criteria for low-energy operation across natural anatomical variations in the mechanical properties of human skin. A wireless, skin-conformable haptic interface, integrating an array of these bistable transducers, serves as a high-density channel capable of rendering input from smartphone-based 3D scanning and inertial sensors. Demonstrations of this system include sensory substitution designed to improve the quality of life for patients with visual and proprioceptive impairments.

This study investigates the effects of two stimulation modalities (stretch and vibration) on natural touch sensation on the volar forearm. The skin-textile interaction was implemented by scanning three textures across the left forearm. The resulting skin displacements were recorded by the digital image correlation technique to capture the information imparted by the textures. The texture recordings were used to create three playback modes (stretch, vibration, and both), which were reproduced on the right forearm. Two psychophysical experiments compared the texture scans to rendered texture playbacks. The first experiment used a matching task and found that to maximize perceptual realism, i.e., similarity to a physical reference, subjects preferred the rendered texture to have a playback intensity of 1X - 2X higher on DC components (stretch), and 1X - 3.5X higher on AC components (vibration), varying across textures. The second experiment elicited similarity ratings between the texture scans and playbacks and showed that a combination of stretch and vibration was required to create differentiated texture sensations. However, the intensity amplification and use of two stimuli were still insufficient to create fully realistic texture sensations. We conclude that mechanisms beyond single-site uniaxial stimuli are needed to reproduce realistic textural sensations.

Electroadhesion (EA) is a promising technology with potential applications in robotics, automation, space missions, textiles, tactile displays, and some other fields where efficient and versatile adhesion is required. However, a comprehensive understanding of the physics behind it is lacking due to the limited development of theoretical models and insufficient experimental data to validate them. This article proposes a new and systematic approach based on electrical impedance measurements to infer the electrostatic forces between two dielectric materials under EA. The proposed approach is applied to tactile displays, where skin and voltage‐induced touchscreen impedances are measured and subtracted from the total impedance to obtain the remaining impedance to estimate the electrostatic forces between the finger and the touchscreen. This approach also marks the first instance of experimental estimation of the average air gap thickness between a human finger and a voltage‐induced capacitive touchscreen. Moreover, the effect of electrode polarization impedance on EA is investigated. Precise measurements of electrical impedances confirm that electrode polarization impedance exists in parallel with the impedance of the air gap, particularly at low frequencies, giving rise to the commonly observed charge leakage phenomenon in EA.

We present PixeLite, a novel haptic device that produces distributed lateral forces on the fingerpad. PixeLite is 0.15 mm thick, weighs 1.00 g, and consists of a 4×4 array of electroadhesive brakes ("pucks") that are each 1.5 mm in diameter and spaced 2.5 mm apart. The array is worn on the fingertip and slid across an electrically grounded countersurface. It can produce perceivable excitation up to 500 Hz. When a puck is activated at 150 V at 5 Hz, friction variation against the countersurface causes displacements of 627 ± 59 μm. The displacement amplitude decreases as frequency increases, and at 150 Hz is 47 ± 6 μm. The stiffness of the finger, however, causes a substantial amount of mechanical puck-to-puck coupling, which limits the ability of the array to create spatially localized and distributed effects. A first psychophysical experiment showed that PixeLite's sensations can be localized to an area of about 30% of the total array area. A second experiment, however, showed that exciting neighboring pucks out of phase with one another in a checkerboard pattern did not generate perceived relative motion. Instead, mechanical coupling dominates the motion, resulting in a single frequency felt by the bulk of the finger.

Electroadhesive surface haptic touchscreens can help augment user experiences by providing tactile effects. The electrode layout in current commercialized designs has separated electrodes for the sensing and actuating functions. During regular use, it is observed that fingerprint residue preferentially deposits on the actuating electrodes far more than the sensing electrodes, which makes the underlying electrode pattern apparent and is highly undesirable for touchscreen users. To address this issue, various physical phenomena (electrohydrodynamic deformation, capillary bridge stabilization, electrowetting, and electrophoretic deposition) are investigated to understand the mechanism. Through experimentation, multiphysics modeling, and surface characterization, it is found that the root cause can be attributed to two mechanisms occurring in the actuating regions: 1) electrohydrodynamic deformation of sebum droplets attached to the finger valleys leading to the formation of additional capillary bridges and residual droplets on the screen surface after their rupture, and 2) electric field-induced stabilization of sebum capillary bridges existing between the finger ridges and the screen, leading to the coalescence and formation of larger-sized droplets. The developed model can then be used to address the issue during the screen design process. An example of using the model to explore the impact of changes in screen oleophobicity is shown.

As the number of applications for tactile feedback technology rapidly increases, so too does the need for efficient, flexible, and extensible representations of virtual textures. The previously introduced Single-Pitch Texel rendering algorithm offers designers the ability to produce textures with perceptually wide-band spectral characteristics while requiring very few input parameters. This paper expands on the capabilities of the rendering algorithm. Diverse families of fine textures, with widely varied spectral characteristics, were shown to be rendered reliably using the Texel algorithm. Furthermore, by leveraging an assistive algorithm, subjects were shown to consistently navigate the Texel parameter space in a matching task. Finally, a psychophysical study was conducted to demonstrate the rendering algorithm's resilience to spectral quantization, further reducing the data required to represent a virtual texture.