Page 65 - Read Online

Page 4 of 15 Zhong et al. Soft Sci. 2025, 5, 3 https://dx.doi.org/10.20517/ss.2024.52

swing arm. The stepper drives the swing arm to oscillate through commands sent by the microcontroller.

Construction of the test apparatus for testing accuracy and resolution of the DCLS sensor
The test apparatus was established, comprising a precision stepper (B6-BYGH156), a bending mechanism,
and a microcontroller (STM32F103RCT6, STMicroelectronics, Italy). The bending mechanism,
manufactured via 3D printing technology (A8s, JG MAKER, China), was designed with interchangeable
components to facilitate various bending radii. This feature enables the assessment of the angular
measurement accuracy of DCLS sensors across different curvatures. The microcontroller sends pulse signals
to the stepper, allowing for highly accurate rotational movements. This test apparatus was also utilized to
evaluate the angular resolution of the DCLS sensor.

RESULTS AND DISCUSSION
Working principle of the proposed sensor
As shown in Figure 1A, the DCLS bending sensor is composed of a soft optical waveguide, a light-emitting
diode (LED), and a chromatic detector. The waveguide comprises a light guide medium (including a red
layer, a blue layer, and a clear core) and a cladding layer. A broadband visible light source (350-770 nm,
5,700 K) is emitted by the LED at one end, coupled into the clear core and then uniformly refracted in the
light guide medium. The chromatic detector is located at the other end to sense the modulated light. To let
the light totally reflect in the optical waveguide, the light guide medium is fabricated using PDMS
possessing a high refractive index (n ≈ 1.41), while the cladding layer is made of a flexible silicone material
[40]
exhibiting a lower refractive index (n ≈ 1.40). According to Snell’s law , the closer refractive indices of the
clear core and cladding, the greater critical angle. Due to the limitations imposed by the material’s refractive
index, in order to achieve total internal reflection within the light guide medium, it is necessary to minimize
the incident angle of the light source as much as possible. Therefore, we use an LED with a slightly larger
emission area (> 3.5 mm × 3.5 mm) than the cross-sectional area of the clear core. This allows the light
emitted by the LED to enter perpendicularly to the cross-section, achieving the minimum incident angle
(detailed information can be found in Supplementary Texts 1 and 2, Supplementary Figures 1 and 2).

In this study, considering the dimensions of most soft robotic manipulators and aiming to adapt to the
applications in these scenarios, the design approach for the sensor size in this research is to simulate the
existing robotic manipulator dimensions and integrate the optical waveguide structure within it. This
integrated design and manufacturing method exhibit high integrability, avoiding the complex assembly
steps and assembly errors associated with traditional sensors that need to be assembled onto soft robots.
This method also saves space and reduces the size of the soft robot, which is fully applicable to most soft
robotic manipulators. Moreover, since the core components of the DCLS sensor include only micro LED
lights and on-chip chromatic detectors, it not only has significantly lower manufacturing costs compared to
traditional bending sensors but also offers a smaller size advantage. Theoretically, while maintaining
existing functionalities, the entire system’s size can be reduced to sub-centimeter levels, providing
possibilities for developing more finely detailed and microscale soft robots in the future.

The sensing principle of the proposed sensor is illustrated in Figure 2. When the DCLS sensor is in a
straight configuration, the input light propagates through the light guide medium via total internal
reflection [Figure 2A]. However, when the DCLS sensor undergoes bending along a rotational axis (where R
represents the bending radius and d signifies half the thickness of the sensor), disparate stretching occurs
across the upper and lower surfaces, as illustrated in Figure 2B. The stretching difference results in distinct
optical path lengths within the red and blue layers. The light intensity attenuation follows the Beer-
[41]
Lambert’s law :