PaperlayerView exact revision on arXiv

Source-linked AI summary

Hall Effect-Based Tactile Force Detection Sensor for Robot-Assisted Minimally Invasive Surgery

Charles DeLorey, Michael Brockdorff, Adam Metcalf, Benjamin Calmé, Pietro Valdastri, Dominic Jones

arXiv:2610.10346v1cs.RO

TL;DR

Remote teleoperation removes surgeons’ natural tactile perception, and local force measurement at the grasper–tissue interface remains unresolved. This paper presents a compact, three-taxel Hall-effect sensor mounted on a surgical grasper jaw. Tests demonstrate force sensing and slip detection, including on ex vivo porcine tissue.

  • Problem

    Remote teleoperation removes natural tactile perception, while reliable local force measurement at the grasper–tissue interface remains unresolved.

    Source: §I · p. 1

  • Method

    A three-taxel Hall-effect array embedded in a deformable elastomer layer senses force through magnet displacement and is mounted on the grasper jaw.

    Source: §I · p. 2, B. Sensing principle · p. 2

  • Results

    74.49 kPa maximum retraction pressure was detected, and ex vivo porcine tests demonstrated grasp-force sensing and slip detection.

    Source: §I · p. 2, §VI · p. 8

  • Takeaways & Limitations

    The sensor demonstrates multi-axis tactile sensing at millimeter scale within surgical-instrument constraints.

    Source: §VI · p. 8

    The sensor adds to the surgical instrument footprint.

    Source: §V · p. 7

Abstract

from arXiv · show

Robotic minimally invasive surgery has revolutionized surgical practice. However, as the surgeons are mechanically separated from the surgical tools, loss of tactile feedback occurs. We present a compact force sensor with three, millimeter scale sensing elements (taxels) that can be affixed to the grasping face of a robotic surgical tool. The sensor features a miniaturized design using Hall effect sensors and magnets embedded in a deformable elastomer contact layer. Sensor calibration is achieved with a 6 degree of freedom (DoF) parallel robot. Each taxel can detect applied normal and shear forces with average RMSE of 2.133 kPa. The calibrated sensor was validated on phantom and ex vivo porcine tissue, demonstrating tactile sensing during retraction as well as slip. This work introduces a small-scale Hall effect based tactile sensor, enabling sensing for force-sensitive robotic surgical applications.

I. INTRODUCTION

RAMIS removes natural tactile perception at the tool tip, while local force measurement at the grasper–tissue interface remains unresolved. The paper introduces an on-jaw Hall-effect sensor to observe contact forces and slip at that interface.

I. INTRODUCTION

  • Excessive grasping can crush tissue and cause ischemia, while insufficient force risks slip and abrasion; remote teleoperation leaves surgeons to infer forces visually.Traditional open surgery and manual laparoscopy rely on tactile sensations through fingertips or tool linkages, whereas remote RAMIS decouples the surgeon from instrument–tissue interaction.Source: §I · p. 1
  • Reliable local force measurement at the grasper–tissue interface remains unresolved; existing approaches estimate forces indirectly or measure bulk interaction without tissue compression.The da Vinci 5 added on-instrument force sensing and visuohaptic feedback, but reports bulk kinesthetic force rather than contact forces experienced by tissue.Source: §I · p. 1
  • On-jaw sensing directly accesses local contact mechanics, including contact pressure, distributed shear, and tissue-dependent friction under variable surgical conditions.Other candidate locations include proximal actuators, the shaft, trocar, and wrist, each providing indirect, global, decoupled, or back-projected measurements.Source: §I · p. 1
  • Related surgical-grasper sensors trade bandwidth, conformability, spatial resolution, or electromagnetic immunity against drift, environmental sensitivity, integration complexity, or cost.Hall-effect approaches offer compact packaging and high sampling rates but require compensation for stray-field and orientation biases,,.Source: §I · pp. 1–2
  • Prior slip approaches offer limited signals or scale: thermal sensing is binary, optics measures displacement, and distributed-force sensing has been demonstrated only in an oversized model.Localized tactile slip sensing directly on the jaw at a scale suitable for RAMIS graspers had not yet been demonstrated, to the authors’ knowledge.Source: §I · p. 2
  • A replaceable on-jaw Hall-effect magnetic sensor provides 3-DoF force and slip observability at the tissue contact interface.The introduction also reports validation on silicone phantoms and ex vivo tissue, with local contact pressures, taxel-wise shear progression during incipient slip, and live force profiles during da Vinci teleoperation.Source: §I · p. 2

II. METHODS

The design targets small-scale multi-axis force sensing across the grasper jaw. Three taxels balance coverage of the grasping surface with the instrument’s physical and manufacturing constraints.

A. Sensor design requirements

  • Three 3-DoF taxels form a miniaturized array small enough to affix to a robotic surgical instrument’s grasping jaw.The paper characterizes the device as small, lightweight, and inexpensive, with multiple taxels enabling directional slip sensing across the grasping face.Source: A. Sensor design requirements · p. 2
  • Multiple taxels cover the grasping surface because tissue rarely spans the full jaw, enabling local pressure measurement and cascading pressure changes that indicate slip.The design maximizes taxel count within the available space, arranging three along the jaw’s principal axis.Source: A. Sensor design requirements · p. 2
  • The design targets normal and shear forces within 0–2 N, 0.1 N resolution, tunable force range, a small footprint, and low-cost scalable manufacture.The paper contrasts its sensing density with commercial sensors, described as bulky and costly.Source: A. Sensor design requirements · p. 2

B. Sensing principle

The sensor infers applied force from magnet displacement relative to Hall sensors, which changes the measured magnetic field. Its three-taxel elastomer skin and electronics are packaged directly on a grasper jaw.

B. Sensing principle

  • Applied pressure or shear displaces embedded magnets relative to Hall sensors, changing magnetic flux density and the sensors’ voltage output.The onboard sensor converts voltage changes into magnetic-field values for force sensing.Source: B. Sensing principle · p. 2
  • Each taxel uses a Hall-sensor pair for force reconstruction, while a proximal pair compensates background stray fields, leaving three taxels in the design..Source: B. Sensing principle · p. 2
  • The sensor’s z-axis is normal to its face, y runs along the three-taxel axis, and x lies perpendicular to y in the sensor face.Source: B. Sensing principle · p. 2

C. Sensor development

  • The skin is made by placing small magnets in a mold, casting and curing Ecoflex 00-30 around them, then removing the cured elastomer.The magnet-bearing skin is affixed to a PCB housing and mounted to the grasper jaw; the magnet-to-sensor distance is 0.7 mm.Source: Figure 3 · p. 3, C. Sensor development · p. 3, Figure 4 · p. 3
  • The assembled sensor measures 9.3×6.5×19.5 mm and weighs 2.06 g; each taxel measures 2.50×1.20×2.50 mm.Rigid NdFeB magnets provide a measurable magnetic-field signal at the small scale.Source: C. Sensor development · p. 3
  • Accounting for sensor thickness reduces the grasper’s maximum open jaw angle from 90°; covering the existing teeth and fenestration makes the sensor face the contact area (107 mm2), sampled at 500 Hz.Source: C. Sensor development · p. 3

D. Sensor calibration

The sensor was calibrated with controlled multi-axis indentations and a neural-network mapping from magnetic measurements to force. Angular testing found a maximum no-contact force-signal deviation of 0.765 kPa across the tested orientations.

D. Sensor calibration

  • Calibration used a 6-axis load cell and parallel robot to apply indentations from 0° to 45° off-normal, to a maximum depth of 0.5 mm.The calibration directions were distributed over 21 locations and sampled at 150 Hz after synchronization and downsampling.Source: D. Sensor calibration · pp. 3–4
  • A feed-forward neural network mapped each taxel’s paired Hall-sensor measurements to three-axis force, with pressure calculated using the 5 mm indenter area.The model used a single 25-neuron hidden layer with ReLU activation and Bayesian regularization; calibration was limited by the load cell’s 0.198 kPa resolution.Source: D. Sensor calibration · p. 4, Table II · p. 4
  • Figure 6 shows hysteresis testing in the top panel and creep testing in the bottom panel, with both tests performed on each taxel.Source: Figure 6 · p. 4
  • 0.765 kPa was the maximum no-contact force-signal deviation during sensor rotations from −40° to +40° on the robot’s first and second joints.The angular evaluation tested whether stray-field compensation limited orientation effects.Source: D. Sensor calibration · p. 4

E. Characterization

Characterization examined magnetic interaction among taxels, magnet-size sensitivity and crosstalk, and the effect of indenter-size mismatch on calibration. The reported tests identify both cross-taxel signal changes and substantial errors when calibration and test contact sizes differ.

E. Characterization

  • FEM simulation found limited magnetic interaction between taxel magnets during displacement, with adjacent-taxel field changes reaching at most 26% in the reported cases.The simulation considered submillimeter N50 neodymium magnets; the taxels use a North-South-North arrangement that draws fields closer to the magnet surface plane.Source: E. Characterization · pp. 4–5
  • Increasing magnet size raised Y- and Z-axis sensitivity, while crosstalk generally remained in the reported range of 12.8%–18.3%.Sensitivity values were 4.15 mT, 13.42 mT, and 15.55 mT for the 1x, 1.5x, and 2x magnets, respectively; the selected size was described as suitable for sensitivity and crosstalk bounds.Source: E. Characterization · p. 5
  • Taxel 1 showed the largest measured crosstalk, at 8.1% from Taxel 2 and 9.9% from Taxel 3 during Z-axis indentation tests.The tests used a five-point trajectory on the indented taxel.Source: E. Characterization · p. 5, Table V · p. 5
  • ±18 kPa errors occurred at the extremes when calibration and test indenter sizes were mismatched, with bias sign reversing according to the direction of the size mismatch.Five indenter sizes were tested; matched-size calibration and tests had RMSE under 1.5 kPa.Source: E. Characterization · p. 5

A. Robotic Evaluation

The in-situ evaluation mounted the sensor on a da Vinci grasper and tested grasping, retraction, and slip. The paper focuses on low-frequency slip and uses changes in each taxel’s shear-to-normal force ratio as an indicator.

A. Robotic Evaluation

  • A da Vinci Research Kit grasper carried the sensor on one jaw, while a bench-mounted 6-DoF load cell provided a reference for bulk material response.A toothed, non-fenestrated opposing jaw was used so pressure could be monitored across the sensing surface rather than concentrated around the perimeter.Source: A. Robotic Evaluation · p. 5, Figure 9 · p. 6
  • For slip tests, silicone and tissue samples were grasped and retracted until slip was observed; lubricant was applied to silicone to emulate soft, wet tissue conditions.The pull force was treated as dominating frictional force in this test setup.Source: A. Robotic Evaluation · p. 5
  • Low-frequency slip was evaluated using each taxel’s FyCoF, the ratio of y-axis shear force to z-axis normal force; a sharp change indicates a slip event.The study focused on low-frequency slip characteristic of soft-tissue interactions.Source: A. Robotic Evaluation · p. 5

B. Experimental Samples

Retraction and slip were assessed on silicone simulants, and samples of ex vivo porcine ileum were also prepared for testing. Silicone retraction produced repeatable multi-axis forces, while slip appeared sequentially across taxels.

B. Experimental Samples

  • Silicone samples measured 100 × 20 × 3 mm and included two fabric layers as a strain-limiting layer to mimic organic tissue.The samples were fixed vertically, grasped at the upper end, and pulled by the robotic instrument.Source: 2) Ex vivo tissue · p. 6

2) Ex vivo tissue:

  • Ex vivo porcine ileum samples were cut to roughly 40 × 5 × 2 mm and stored in phosphate-buffered saline to maintain hydration.Source: 2) Ex vivo tissue · p. 6

A. Silicone

  • Silicone retraction produced X-, Y-, and Z-axis forces with cyclic tugging; peak pressures were 27.5 kPa in Y and 91.2 kPa in Z across five pulls.The Y- and Z-axis measurements had standard deviations below 0.28 kPa across pulls.Source: A. Silicone · p. 6
  • Slip began on Taxel 1 at approximately 14.25 kPa Y-axis shear force and propagated distally, with peak Y-axis forces of approximately −31.59, −27.41, and −8.14 kPa on Taxels 1, 2, and 3.The reported onset time was approximately 4.2 seconds.Source: A. Silicone · p. 6

B. Tissue

Tissue retraction produced lower peak pressures than silicone, while slip experiments showed measurable but irregular forces and variation across taxels.

B. Tissue

  • 11.9 kPa in Y and 15.3 kPa in Z were tissue retraction maxima, around 50% below silicone.Tissue maxima had standard deviations of 0.038 kPa in Y and 0.081 kPa in Z.Source: B. Tissue · p. 6, Figure 11 · p. 7
  • ~2.5x increase in max pull force occurred from 10 mm to 15 mm retraction (29.592 kPa [3.974 kPa] versus 74.49 kPa [11.939 kPa]).The authors report that the sensor maintained a solid grasp while providing shear-force information.Source: B. Tissue · p. 6
  • Tissue slipped within the jaws but did not leave the grasper, and load-cell Z force returned to zero at the home position.Source: B. Tissue · p. 6, Figure 11 · p. 7
  • Tissue slip produced more irregular, gradual forces than silicone, and repeated responses indicated sensitivity to biologically relevant ex vivo loading.Source: B. Tissue · pp. 6–7
  • Across 11 slip samples, Taxel 1 CV was <0.64 and Taxel 2 CV was >1.6; Y forces were most reproducible and Z forces most variable.The authors attribute the asymmetry to the scissor-grasp geometry and variable force distribution during slip.Source: B. Tissue · p. 7

C. Freeform tissue grasping

Freeform porcine-tissue handling produced stable multi-taxel force traces, while discussion results describe the sensor’s operating scope and constraints. Tissue slip responses varied, and the sensorized jaw altered grasp robustness.

C. Freeform tissue grasping

V. DISCUSSION

  • The three-taxel sensor fits a surgical grasper jaw, resolves at least 6% of maximum pressure, and targets 0–255 kPa RAMIS forces.Magnet choice, sensor distance, and elastomer can tune sensitivity, but the sensor adds to the instrument footprint.Source: §V · p. 7
  • The flat, smooth elastomer face permits more slip than the unmodified toothed jaw, resulting in less strong and robust grasps.Source: §V · p. 7
  • Tissue traits and contact conditions vary grasp outcomes, while principal-axis retraction and slip tests simplify clinical use.The passage identifies thickness, hydration, elasticity, contact area, and pull angle or plane as relevant factors.Source: §V · p. 7
  • The yCoF metric displayed slip in silicone and tissue, but tissue yCoF values varied greatly and indicated non-uniform slip interactions.The authors suggest inconsistent contact, variable tissue properties, and stick-slip friction may contribute, motivating study across tissue types.Source: §V · pp. 7–8

VI. CONCLUSIONS

The miniaturized Hall-effect sensor integrates into a robotic grasper and demonstrated grasp-force sensing and slip detection on ex vivo porcine tissue. Its architecture is presented as customizable and scalable, while the current instrument modification and future calibration needs remain limitations.

VI. CONCLUSIONS

  • The Hall-effect sensor was calibrated with ground-truth force data and demonstrated grasp-force sensing and slip detection on ex vivo porcine tissue.It is integrated directly into the grasping surface of a robotic surgical instrument.Source: §VI · p. 8
  • The sensing architecture supports customizable taxel configurations and is described as scalable for high-volume production as a disposable add-on sensor.Source: §VI · p. 8
  • Further work could examine non-normal angle indentations for more detailed calibration and use more intricate magnetic profiles to improve sensor resolution and sensitivity.Source: §VI · p. 8
  • The current iteration sensorizes one grasping surface, a modification that likely altered the instrument’s original grasping capabilities.Further development would evaluate the modified tool grasp to minimize performance interference; sterilization and biocompatible materials also remain considerations for clinical viability.Source: §VI · p. 8
Loading 2610.10346v1…