Polymer-Based Self-Calibrated Optical Fiber Tactile Sensor

Wentao Chen, Youcan Yan, Zeqing Zhang, Lei Yang, Jia Pan

TL;DR

This paper introduces a novel polymer-based optical fiber tactile sensor that can simultaneously measure normal and shear forces, and crucially, the size of the object it's interacting with. By leveraging the anisotropic response of a two-layered woven optical fiber structure, the sensor achieves self-calibration, decoupling these measurements without relying on large datasets or complex machine learning models.

What Problem Does This Paper Address?

Robotic manipulation often requires a nuanced understanding of contact, akin to human touch. While existing tactile sensors can detect forces, accurately decoupling normal and shear forces has been a challenge. Furthermore, a significant limitation has been the failure to account for the size of the object being touched, which can significantly impact force calibration. This oversight can lead to inaccurate force readings, especially when dealing with objects of varying geometries. The authors aim to bridge this gap by developing a sensor that not only measures forces but also incorporates object size into its calibration, leading to more robust and accurate tactile perception for robots.

Methods

The core of the sensor's design lies in a two-layered structure of soft optical fibers embedded within a compliant elastomer. The key innovation is the anisotropic arrangement of these fibers: the upper layer features U-shaped fibers optimized for shear force detection, while the lower layer has straight fibers primarily sensitive to normal forces. This structural anisotropy, combined with the inherent property of optical fibers where light power loss correlates with deformation, allows for the decoupling of different force components and object size.

The sensor utilizes infrared (IR) LEDs for light emission and photodiodes (PDs) for detection. The manufacturing process involves topographic molding of the elastomer, PCB layer fabrication with integrated LEDs and PDs, and careful assembly.

A linear self-calibration algorithm is proposed, which operates in two steps:

  1. Indentation Depth and Size Calibration: This step uses the signals from the lower layer fibers and a bottom PD to establish a relationship between light intensity changes and indentation depth/object radius.
  2. Force Calibration: This step builds upon the first, using all six optical fibers to calibrate the relationship between light intensity changes and applied normal and shear forces, taking into account the previously determined indentation factors.

Datasets

The paper describes calibration and testing procedures using a controlled experimental setup. They collected data by applying forces with cylindrical indenters of varying diameters (5mm to 12mm) at different positions and orientations. The ground truth force data was obtained from a Force/Torque sensor. No specific named dataset is mentioned as being used for training or evaluation; rather, the authors generated their own calibration and test datasets through their experimental setup.

Key Insights and Ideas

  • Anisotropic Fiber Arrangement is Key: The deliberate arrangement of optical fibers in different orientations and shapes (U-shaped vs. straight) within distinct layers is crucial for decoupling normal, shear forces, and object size.
  • Object Size Matters for Force Calibration: Ignoring object size leads to inaccuracies. This work demonstrates that incorporating object size into the calibration process significantly improves force measurement reliability.
  • Linear Self-Calibration is Sufficient: For the tested range, a linear calibration model proves effective in simultaneously estimating forces and object size, avoiding the need for complex, data-hungry machine learning approaches.
  • Optical Fiber Properties Enable Sensing: The fundamental principle of light power loss in optical fibers due to deformation is effectively harnessed for tactile sensing.

Key Results

The tactile sensor demonstrated good performance in reconstructing force vectors and measuring object size:

  • Force Reconstruction Accuracy:
    • Normal force (Z-axis): Average accuracy of 0.15N.
  • Shear force (X-axis): Average accuracy of 0.17N.
  • Shear force (Y-axis): Average accuracy of 0.18N.
  • These results were achieved within a sensing range of 0-2N in all directions.
  • Object Size Measurement Accuracy:
    • Average accuracy of 0.4mm for indenter diameter.
  • This was within the tested indenter diameter range of 5-12mm.

The paper also shows that the anisotropic responses of the upper and lower layer fibers are distinct, with upper fibers being more sensitive to shear and lower fibers to normal forces, and that the straight fibers in the lower layer show sensitivity to both indentation depth and indenter diameter.

Limitations and Future Work

  • Limited Sensing Range: While the accuracy is good within the 0-2N range, the paper acknowledges that the linear calibration might become less effective at higher force magnitudes due to increased nonlinearity in elastomer deformation.
  • Peripheral Component Size: The current sensor design is somewhat limited by the size of the peripheral electronic components (LEDs, PDs, and PCB).
  • Future Work Suggestions:
    • More Layers/Anisotropies: Adding more layers or exploring different fiber arrangements could enable the sensing of additional tactile modalities.
  • Optimized Fiber Arrangement: Further optimization of fiber placement could enhance anisotropy and sensing capabilities.
  • Conformability and Wearability: The sensor's conformable nature and potential for integration into wearable devices are highlighted as promising avenues for future development.

Summary updated

Abstract

Human skin can accurately sense the self-decoupled normal and shear forces when in contact with objects of different sizes. Although there exist many soft and conformable tactile sensors on robotic applications able to decouple the normal force and shear forces, the impact of the size of object in contact on the force calibration model has been commonly ignored. Here, using the principle that contact force can be derived from the light power loss in the soft optical fiber core, we present a soft tactile sensor that decouples normal and shear forces and calibrates the measurement results based on the object size, by designing a two-layered weaved polymer-based optical fiber anisotropic structure embedded in a soft elastomer. Based on the anisotropic response of optical fibers, we developed a linear calibration algorithm to simultaneously measure the size of the contact object and the decoupled normal and shear forces calibrated the object size. By calibrating the sensor at the robotic arm tip, we show that robots can reconstruct the force vector at an average accuracy of 0.15N for normal forces, 0.17N for shear forces in X-axis, and 0.18N for shear forces in Y-axis, within the sensing range of 0-2N in all directions, and the average accuracy of object size measurement of 0.4mm, within the test indenter diameter range of 5-12mm.
BibTeX
@inproceedings{chen2023polymer,
  author = {Wentao Chen and Youcan Yan and Zeqing Zhang and Lei Yang and Jia Pan},
  title = {{Polymer-Based Self-Calibrated Optical Fiber Tactile Sensor}},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
  year = {2023},
  url = {https://openreview.net/forum?id=RaXxIBvTsA},
}
Figure