Fiber Optic Shape Sensing for Soft Robotics

Abstract

While soft material actuators can undergo large deformations to execute very complex motions, what is critically lacking in soft material robotic systems is the ability to collect high-resolution shape information for sophisticated functions such as environmental mapping, collision detection, and full state feedback control. This work explores the potential of a nearly commercial fiber optic shape sensor (FOSS) and presents the first demonstrations of a monolithic, multicore FOSS integrated into the structure of a fiber-reinforced soft actuator. In this pilot study, we report an open loop sensorized soft actuator capable of submillimeter position feedback that can detect the soft actuator's shape, environmental shapes, collision locations, and material stiffness properties.

Keywords: fiber optic shape sensor; soft actuator; soft robotics; soft sensor.

FIG. 1.

(A) Generalized OFDR optical network (B) Illustration of helical cores along a length of fiber with the inset image depicting the sensing triad—labeled 1, 2, and 3—and the core fiber—labeled 0. (C) Typical four-core strain response of the FOSS under pure bending, where the outer cores show sinusoidal strain centered around 0, while the center core remains neutral. (D) Typical four-core strain response of the FOSS under bending and twist, where the center core strain remains neutral, while the average of the outer cores is shifted upward or downward. FOSS, fiber optic shape sensing; OD, outside diameter; OFDR, optical frequency domain reflectometry; SEM, scanning electron microscope. Color images are available online.

FIG. 2.

Schematic diagram highlighting the fabrication process for a soft fiber-reinforced bending actuator with integrated lumens. (A) The first molding step using a 3D printed three-part mold to define the exterior shape of the rubber body around a half-round steel rod. (B) The strain limiting layer (woven fiberglass) is attached to the flat face of the actuator. (C) Fiber reinforcements (Kevlar fiber) are wound along the entire length of the actuator. (D) The second molding step, the entire actuator is encapsulated in a 2.0 mm thick layer of silicone to anchor all fiber reinforcements and to create lumens with the 3.175 mm diameter steel rods that run parallel to the long axis of the half round rod. (E) All the steel rods are removed and both ends of the actuator are capped with a connection point on one end for pressurized fluid. (F) Fabricated soft actuator measuring 16 cm in length. 3D, three-dimensional. Color images are available online.

FIG. 3.

(A) Path of the FOSS when it is passed through the actuator's lumens. (B) U-shape of the sensor as loops around the end to return through the other lumen. We assume that the center line of the FOSS lies in a plane—sensor plane. (C) Exploded view of the rigid plastic cap that is attached to the end of the sensor to reinforce the U-shape in the sensor plane. Color images are available online.

FIG. 4.

Soft robot tip position and cap position as the tip of the actuator is dragged along a surface of a sinusoid shape. Color images are available online.

FIG. 5.

Experimental setup for measuring the twist at the tip of the actuator, while the base of the actuator is anchored. The tip of the actuator was rotated in increments of 5° to a maximum of −90° and +90°. Color images are available online.

FIG. 6.

Top view of the platform used for the linear translation, contact detection, and shape detection experiments. Color images are available online.

FIG. 7.

Experimental setup for tracking the tip of the actuator when the actuator-sensor combination was moved 200 mm along a linear path on a motorized linear stage. Color images are available online.

FIG. 8.

Overlays of the actuator-sensor combination at different pressures, illustrating the experimental setup for tracking the tip of the actuator at multiple points on the optical breadboard that lie closest to its free deflection path. The optical breadboard served as the ground truth for evaluating the FOSS data. Color images are available online.

FIG. 9.

(A) Top view of the collision detection experiment with the location of the aluminum standoff obstacle at 18, 58, and 98 mm relative to the base of the actuator. (B) Depicts the instant the soft actuator contacts the obstacle. The shape of the sensor is recorded as FOSS state 0. (C) Depicts the deformation of the actuator-sensor combination some distance further down the linear stage. The shape of the sensor is recorded as state 1. (D) Diagram of the shape data from states 0 and 1 overlaid to identify their intersection and thus the location of the obstacle. Color images are available online.

FIG. 10.

(A) Top view of the actuator tip sweeping over a 16 mm amplitude sawtooth surface that was anchored to the optical table. It should be noted that the red arrows highlight the path of the actuator tip as it sweeps over a surface. (B) Overview of all the surface geometries that were evaluated. Color images are available online.

FIG. 11.

Top view of the actuator-sensor combination anchored to an optical table using right angle mounts with the tip of the actuator resting on the surface of a test sample. The inset image shows a close-up of the actuator pressurized to 276 kPa (40 psi) and the tip pressing into the sample material. Color images are available online.

FIG. 12.

(A) Tip angle measured by the FOSS as it was rotated in 5° increments from +90° to −90°. (B) 3D view of the FOSS shape data at three different twist positions—0°, 45°, and 90°. Color images are available online.

FIG. 13.

Tip position of the soft actuator, which is 1.25 m from the FOSS origin, as the actuator traversed 200 mm along the motorized linear stage. Color images are available online.

FIG. 14.

(A) Tip position of the soft actuator as measured by the FOSS compared to the known anchor points in the optical breadboard. (B) Side view and (C) 3D view of the shape data at several different pressures—0, 10, 20, and 40 psi. Color images are available online.

FIG. 15.

Experimental results of the actuator-sensor combination to inspect and reconstruct six different planar shapes, with submillimeter error: (A) flat, (B) 4 mm amplitude sawtooth, (C) 16 mm amplitude sawtooth, (D) convex curve, (E) concave curve, (F) sinusoid. Color images are available online.

FIG. 16.

Experimental results of the actuator-sensor combination to detect the relative difference in stiffness across a range of compliant materials. Color images are available online.