Friction-Scaled Vibrotactile Feedback for Real-Time Slip Detection in Manipulation using Robotic Sixth Finger

TL;DR

This work tackles the lack of natural sensory feedback in extra robotic fingers by encoding friction information and incipient slip into vibrotactile cues delivered to the user. The authors implement a friction-scaled haptic feedback system guided by the second derivative of tangential force $(d^2f_t/dt^2)$ with a threshold of $0.3~\mathrm{N}/\mathrm{s^2}$ and a slip-ratio difference criterion, mapping slip onset and surface friction to vibrotactile cues. In a 2AFC psychophysics experiment across three friction levels, they report an overall slip-friction discrimination accuracy of $94.53\% \pm 3.05\%$, with peak tangential forces modulated by surface friction and two reliable slip-detection criteria driving real-time feedback. The study demonstrates the feasibility of tactile feedback for SRLs to enhance grip stability and points toward future automatic grip-force regulation and texture sensing, with implications for rehabilitation and assistive robotics. This friction-aware vibrotactile approach offers a noninvasive, scalable path to more natural and autonomous control of extra limbs in daily tasks.

Abstract

The integration of extra-robotic limbs/fingers to enhance and expand motor skills, particularly for grasping and manipulation, possesses significant challenges. The grasping performance of existing limbs/fingers is far inferior to that of human hands. Human hands can detect onset of slip through tactile feedback originating from tactile receptors during the grasping process, enabling precise and automatic regulation of grip force. The frictional information is perceived by humans depending upon slip happening between finger and object. Enhancing this capability in extra-robotic limbs or fingers used by humans is challenging. To address this challenge, this paper introduces novel approach to communicate frictional information to users through encoded vibrotactile cues. These cues are conveyed on onset of incipient slip thus allowing users to perceive friction and ultimately use this information to increase force to avoid dropping of object. In a 2-alternative forced-choice protocol, participants gripped and lifted a glass under three different frictional conditions, applying a normal force of 3.5 N. After reaching this force, glass was gradually released to induce slip. During this slipping phase, vibrations scaled according to static coefficient of friction were presented to users, reflecting frictional conditions. The results suggested an accuracy of 94.53 p/m 3.05 (mean p/mSD) in perceiving frictional information upon lifting objects with varying friction. The results indicate effectiveness of using vibrotactile feedback for sensory feedback, allowing users of extra-robotic limbs or fingers to perceive frictional information. This enables them to assess surface properties and adjust grip force according to frictional conditions, enhancing their ability to grasp, manipulate objects more effectively.

Figures (8)

• Figure 1: Instrumented Robotic Sixth Finger: Participant performing griping and lifting task using the instrumented robotic sixth finger
• Figure 2: Schematic of the whole control process after getting force feedback from the sensor and sensing haptic feedback to the user
• Figure 3: Experimental setup and protocol: A, Participant wearing robotic sixth-finger instrumented with an ATI-Nano17 Force-torque sensor, an armband with an ERM vibration motor, and an eye shield B, GUI designed to guide the experimenter during trials where the sixth finger grips a glass with three different friction levels. The GUI prompts the experimenter to initiate the grip and, after reaching a threshold normal force of 3.5N, lift the glass. Release begins after lift-off. Real-time bar charts display force/torque levels to ensure target forces are met and to prevent damage during process C, the schematic of the experimental sequence D, schematic illustration of the time course of presentation and evaluation of the stimulus pairs by subjects touching the friction modulation device (H vs. M and H vs. L denote pairs of stimuli where H is high friction; L is low friction; M is medium friction). Glass is gripped for the L condition and sandpaper with different grit no. is gripped for the M and H condition.
• Figure 4: Participants performance: Barplots displaying the mean and standard deviation percentages of correct responses across friction pairs. $****P < = 0.00001$ (Bonferroni corrected). Blue circles show the individual responses for the HvsM condition, green triangles show individual responses for the MvsL condition and red squares show individual responses for the HvsL condition
• Figure 5: Normal force and tangential force traces as a function of time: Each panel represents the average of the $20$ individual trials. Force traces are superimposed, solid lines indicate the normal force and the dotted line indicates the tangential force. The blue lines represent the forces during the high friction condition, the red lines represent the forces during the medium friction condition and the green lines represent the forces during the low friction condition. The inset in each graph represents only the mean tangential forces and the standard deviation synchronized at $0.5N$ for all the $20$ individual trials for all frictional conditions. The solid line indicates the mean and the shaded region indicates the standard deviation.