PaperlayerView exact revision on arXiv

Source-linked AI summary

Remote Surfaces at Your Fingertips: Electrovibration-Based Tactile Feedback for Robot Teleoperation via Touchscreen Interfaces

Alperen Kenan, Juan José García Cárdenas, Adriana Tapus, Paul Bremner, Manuel Giuliani

arXiv:2609.27938v1cs.ROcs.HC

TL;DR

Contact-rich teleoperation needs force awareness, but kinesthetic feedback can be unstable at rigid contacts and visual cues add cognitive demands. This paper maps remote interaction forces to touchscreen electrovibration, and reports faster responses and higher telepresence than visual feedback.

  • Problem

    Kinesthetic feedback can be unstable during stiff contacts and sensitive to communication delays, while visual force cues require an additional sensory channel.

    Source: §I · p. 1

  • Method

    A touchscreen interface maps remote robot-environment interaction forces to electrovibration feedback at the operator’s fingertip, evaluated with 21 participants.

    Source: §I · p. 2

  • Results

    TF participants responded 71.68 ms faster than VF participants (15.35% reduction, p < 0.001, d = 0.96), and TF yielded 31% higher sense of telepresence than VF (p < 0.001, d = 0.90).

    Source: §VII · p. 7

  • Takeaways & Limitations

    Tactile feedback’s main advantage was higher telepresence and lower response latency rather than reduced workload, with greatest benefits in contact-rich, multi-surface tasks.

    Source: §VII · p. 7, §VIII · p. 8

    The 21-participant study used a simulated robot in a laboratory with non-expert operators, limiting generalisability and direct applicability to industrial practice.

    Source: §VII · p. 7

Abstract

from arXiv · show

Enabling operators to perceive and interact with remote environments naturally is a fundamental challenge in robotic teleoperation. This is especially critical in tasks involving physical interaction, where real-time haptic awareness improves operational safety and effectiveness. Existing kinesthetic haptic feedback methods suffer from instability during rigid surface contacts and remain sensitive to communication delays, while visual cue-based force feedback imposes additional cognitive load and limits sustained situational awareness. This work presents a teleoperation interface that conveys remote surface interactions to the operator through electrovibration-based tactile feedback, enabling naturally mapped force reflection while avoiding the stability issues associated with kinesthetic feedback and the latency limitations of mechanical actuators. A user study (N=21) evaluated interface usability, sense of presence, and operator workload under two force reflection conditions: visual feedback and electrovibration-based tactile feedback. Characterisation experiments further assessed path-following accuracy and response time across both conditions. Results show that tactile feedback significantly reduced response time by 15.35% (p=0.002, d=0.96) and increased the sense of presence by 31% (p<0.001, d=0.90) compared to visual feedback, while imposing comparable workload and usability across both conditions. These findings demonstrate that electrovibration-based tactile feedback is a viable and effective modality for robot teleoperation, improving operator responsiveness and sense of presence in contact-rich manipulation tasks, with direct applicability to safety-critical domains such as nuclear maintenance.

I. INTRODUCTION

Robot teleoperation for contact-rich surface tasks demands precise force control and sustained awareness, while existing feedback modalities impose stability or cognitive-load trade-offs. This paper introduces a touchscreen interface that reflects remote interaction forces through electrovibration at the operator’s fingertip.

I. INTRODUCTION

  • Kinesthetic feedback can become unstable during stiff contacts and sensitive to communication delays, while visual force cues require an additional sensory channel.Tactile feedback is presented as an alternative that conveys interaction forces through the same modality as physical contact.Source: §I · p. 1
  • Electrovibration modulates fingertip friction through an alternating voltage on a conductive screen, and its use for reflecting remote surface interactions in teleoperation remains unexplored.The paper identifies low signal latency and distinguishable tactile sensations as relevant properties.Source: §I · p. 1
  • The interface maps remote robot-environment interaction forces to electrovibration-based tactile feedback through a touchscreen with negligible delay.The stated aims include reducing workload, improving safety and task performance, and enhancing telepresence.Source: §I · pp. 1–2
  • The work implements electrovibration force reflection, designs and evaluates a low-cost non-wearable interface, and studies five operator outcomes with 21 participants.The outcomes include response time, path-following accuracy, telepresence, cognitive load, and usability across two feedback conditions.Source: §I · p. 2

II. RELATED WORK

Prior work motivates tactile feedback for contact-rich teleoperation, while identifying electrovibration as a touchscreen-compatible force-reflection channel not previously applied to robot teleoperation. The study evaluates this interface against visual feedback with 21 participants.

II. RELATED WORK

  • Touchscreens offer direct finger-to-end-effector mapping, but prior work identified absent haptic feedback as a limitation for maintaining situational awareness.Earlier work by the authors found improved efficiency and path-tracking accuracy over a joystick with reduced cognitive load.Source: §II · p. 2
  • Electrovibration modulates electrostatic fingertip friction on a conductive screen and has been studied for rendering distinguishable tactile sensations, but not as a real-time teleoperation force-reflection channel.Related work proposes integrating it with end-effector tactile sensors as a route to low-latency tactile telepresence.Source: §II · p. 2

III. METHODOLOGY

  • The comparative study assessed interface performance, usability, telepresence, and cognitive load through characterization and user-evaluation tasks modeled on contact-based surface manipulation.Scenarios included responding to feedback, following a guided path, swabbing multiple surfaces, and avoiding obstacles.Source: §III · p. 2

A. Participants

  • Twenty-one participants aged 19–37 took part; all used their right hand, and prior robot-control experience ranged from none to regular use.The group included 15 men and 6 women.Source: A. Participants · p. 2

B. Study Environment and Context

The study used a simulated robot and touchscreen interface to compare visual and tactile feedback across characterization and surface-interaction evaluation tasks. Task order and rendered environments were randomized across participants.

B. Study Environment and Context

  • The teleoperated robot and environment were simulated in Unity, with a simulated UR10 arm used throughout the experiments.Participants interacted with the interface at the Bristol Robotics Laboratory.Source: B. Study Environment and Context · pp. 2–3
  • Characterization compared visual feedback via a side bar graph with tactile feedback and tested path following and response time under identical camera feeds.The path-following task used unseen sine-wave trajectories; the response-time task asked participants to lift a finger on feedback.Source: Figure 2 · p. 3
  • User-evaluation tasks covered brick-wall defect detection, multi-surface swabbing, and obstacle avoidance, with feedback representing surface textures or obstacle proximity.The multi-surface task used different electrovibration frequencies and amplitudes; colors represented textures in the visual-feedback condition.Source: B. Study Environment and Context · p. 3, Figure 4 · p. 3
  • The touchscreen combined a display, conductive electrovibration layer, and infrared touch-tracking frame in a single integrated assembly.The three layers enabled visual output, tactile feedback delivery, and finger-position tracking.Source: B. Study Environment and Context · p. 3
  • All 21 participants completed both phases and conditions after practice trials, with phases, conditions, paths, and environments randomized to minimize learning effects.Participants were not told the study hypotheses.Source: B. Study Environment and Context · pp. 3–4

C. Subjective Measurements

The study measured workload, usability, and presence using post-condition questionnaires, and collected open-ended feedback. The protocol received institutional ethics approval and included consent and participant data protections.

C. Subjective Measurements

  • After each condition, participants completed NASA-TLX for cognitive load and SUS for usability, alongside a four-item presence questionnaire rated on a five-point Likert scale.Questionnaire responses were converted to a 100-point metric, and participants also provided open-ended feedback.Source: C. Subjective Measurements · p. 4

D. Ethics

  • The University of the West of England ethics committee approved the study, and participants provided signed consent before experiments.Participants could withdraw and request data removal for up to seven days after the study.Source: D. Ethics · p. 4
  • The project repository includes hardware lists, models of printed and laser-cut parts, feedback scripts, and the robot simulation environment.These materials are provided for reproducibility.Source: D. Ethics · p. 4

IV. TECHNICAL DETAILS OF THE TELEPRESENCE IMPLEMENTATION

The interface uses voltage-driven electrovibration to modulate fingertip friction in time with remote surface features. Its setup generates and amplifies the signals while enclosing high-voltage outputs and limiting current for safe operation.

A. Tactile Feedback through Electrovibration

  • Timed alternating voltage increases electrostatic force and fingertip friction, allowing the interface to reproduce remote surface defects through touch.The system maps surface region to voltage amplitude, frequency, and waveform; response time reflects the complete teleoperation loop, whose delays were common to both conditions.Source: A. Tactile Feedback through… · p. 4, Figure 6 · p. 4
  • Unlike mechanical vibration, electrovibration has no moving parts, and the perceived sensation is mechanical rather than electrical.Current through tissue remains below the electrical-stimulation threshold.Source: A. Tactile Feedback through… · pp. 4–5

B. Tactile Feedback Experiment Setup

  • A DRV2667EVM-CT module amplifies an audio-converted analogue reference; the primary tactile signal was a 200 Vpp, 120 Hz square wave.The multi-surface implementation varied signal parameters across surfaces; Fig. 7 shows an example voltage difference between the two ports.Source: B. Tactile Feedback Experiment Setup · p. 5, Figure 7 · p. 5
  • An isolator box enclosed power outputs, and a 100 kΩ resistor limited touchscreen-line current to 1 mA in an unexpected short circuit.Shrouded connectors prevented unintended contact, while a wristband resistor reduced current through the fingertip to approximately 0.2 mA.Source: B. Tactile Feedback Experiment Setup · p. 5, Figure 8 · p. 5

V. HYPOTHESES

The hypotheses predict that electrovibration can convey perceptible force feedback without notable delay, and that tactile feedback will improve response time and telepresence relative to visual cues. They also predict lower cognitive workload under tactile feedback.

H1

  • H1 predicts that electrovibration conveys perceivable force feedback without notable delays.The rationale cites prior evidence that electrovibration can render tactile feedback under optimal operating conditions; participants were to report whether they perceived it clearly.Source: §V · p. 5

H2

  • H2 predicts faster response times with tactile feedback than with visual cue-based force feedback.The rationale contrasts direct fingertip feedback with visual processing and a planned motor response.Source: §V · p. 5

H3

  • H3 predicts a higher sense of telepresence with tactile feedback than with visual feedback.The rationale is that matching feedback to the physical interaction may strengthen presence; visual cues require cognitive conversion, whereas tactile cues reach the fingertip directly.Source: §V · pp. 5–6

H4

  • H4 predicts lower cognitive workload with tactile feedback than with visual feedback.The rationale is that visual force monitoring divides operator attention; questionnaires assess cognitive load, usability, and telepresence.Source: §V · pp. 5–6

VI. RESULTS

The user study had a task-performance phase and an operator-perception phase, with all 21 participants included in both.

VI. RESULTS

  • The user study comprised two phases.One phase measured task performance; the other assessed operator perception of the interface with tactile feedback.Source: §VI · p. 6
  • The first phase measured task performance.The second phase assessed operator perception of the proposed teleoperation interface with tactile feedback.Source: §VI · p. 6
  • All 21 participants completed the study and contributed data to both phases.Source: §VI · p. 6

A. Characterisation Experiment Results

Tactile feedback shortened response time, while path-following accuracy did not differ significantly between tactile and visual feedback conditions.

A. Characterisation Experiment Results

  • Trial measurements were averaged per participant and condition, yielding 21 paired samples for each reported test.Path-following measurements averaged two trials per condition; response-time measurements averaged three.Source: A. Characterisation Experiment Results · p. 6
  • VF and TF mean position errors were 65.96 and 63.84 pixels (p = .76), while maximum errors were 228.20 and 226.00 pixels (p = .92); neither difference was significant.Source: A. Characterisation Experiment Results · p. 6
  • Response time was 15.35% lower with TF than VF, a 71.68 ms reduction (p = .0016, d = 0.79).Mean response times were 395.24 ms for TF and 466.92 ms for VF; Fig. 9 presents the characterisation results as boxplots.Source: Figure 9 · p. 6, A. Characterisation Experiment Results · p. 6

B. User Evaluation Results

Across the 21 responses per condition, tactile feedback increased telepresence, while usability and workload were statistically comparable to visual feedback. Task rankings differed by feedback condition.

B. User Evaluation Results

  • NASA-TLX scores did not differ significantly between VF and TF (24.78 vs. 27.04; p = .57, d = 0.13).Questionnaire scores were normalised to a 100-point scale, with higher NASA-TLX scores indicating higher cognitive load.Source: B. User Evaluation Results · p. 6
  • SUS scores did not differ significantly between VF and TF (79.88 vs. 78.39; p = .60, d = 0.12).Higher SUS scores indicate greater usability.Source: B. User Evaluation Results · p. 6
  • 17.5 points (31%) higher average telepresence score under TF than VF (73.87 vs. 56.36; p < 0.001, d = 0.90).Figure 10 visualises the user-evaluation results as boxplots.Source: B. User Evaluation Results · p. 6, Figure 10 · p. 6
  • Task rankings varied by condition: VF showed no significant differences across tasks (p = .157), while TF did (p = .046), ranking swabbing and defect detection above obstacle avoidance.A rank of 1 indicated the most important task and 3 the least important.Source: Figure 10 · p. 6

VII. DISCUSSION

The discussion identifies faster responses and stronger telepresence—not reduced workload—as tactile feedback’s clearest benefits, especially for contact-rich manipulation. It also describes implementation constraints and the need for real-robot evaluation.

VII. DISCUSSION

  • All 21 participants reported successful tactile perception, but perception varied; the current differential-output setup also requires a wristband that slightly restricts movement.The discussion links perception differences to body grounding and proposes integrating a grounding mechanism.Source: §VII · p. 7
  • TF responses were 71.68 ms faster on average than VF responses, a 15.35% reduction (p < 0.001, d = 0.96).The discussion treats this as support for H2.Source: §VII · p. 7
  • TF yielded 31% higher telepresence than VF (p < 0.001, d = 0.90), with the largest reported gap in feeling direct control of the robot.Direct-control scores were 38.21 for VF and 68.33 for TF.Source: §VII · p. 7
  • NASA-TLX showed no significant workload difference between conditions; unfamiliar electrovibration cues may help explain the comparable demands.Participants also often preferred TF, and some suggested combining tactile and visual feedback.Source: §VII · p. 7
  • TF task rankings favored multi-surface swabbing and defect detection over obstacle avoidance (F(2, 40) = 3.33, p = .046).The discussion relates this pattern to tactile cues not increasing as linearly as visual feedback.Source: §VII · p. 7
  • The discussion attributes the approach’s advantages to low-latency electrical feedback without moving parts, natural tactile-to-tactile mapping, and avoiding kinesthetic instability during stiff contacts.Potential extensions include medical telerobotics, assistive technologies, and remote operations needing enhanced situational awareness.Source: §VII · p. 7
  • Electrovibration requires fingertip motion for sensation, and its perceived friction force does not map one-to-one to environment contact force.These constraints limit feedback during static contact and constrain force fidelity.Source: §VII · p. 7
  • The study used 21 non-expert participants and a simulated robot in a laboratory; real-robot network delay and sensor noise may affect tactile-cue timing and reliability.Future work will test real robots and domain experts, including communication delay and sensor noise.Source: §VII · pp. 7–8

VIII. CONCLUSION

The interface’s electrovibration-based tactile feedback reduced response latency and enhanced telepresence compared with visual feedback, with comparable cognitive load. Its benefits were most pronounced in contact-rich surface interactions, and the authors present it as a practical approach for precision and situational awareness.

VIII. CONCLUSION

  • Characterisation experiments found sufficiently low latency, perceptible tactile feedback for all participants, and faster response times than visual feedback.Source: §VIII · p. 8
  • Tactile feedback significantly enhanced telepresence and reduced operator response latency while imposing comparable cognitive load to visual feedback.Task rankings showed the benefits were most pronounced in contact-rich tasks such as multi-surface swabbing and defect detection.Source: §VIII · p. 8
  • The authors describe the interface as a practical, low-cost approach to precision and situational awareness in surface interaction tasks.They state that perceiving remote interactions through the fingertip helps reduce error risk in hazardous environments.Source: §VIII · p. 8
Loading 2609.27938v1…