Hold 'em and Fold 'em: Towards Human-scale, Feedback-Controlled Soft Origami Robots

TL;DR

This work tackles the dual challenges of proprioceptive feedback control and actuation under human-scale loads in soft robotics. It introduces a fluidic-actuated, one-DOF Kresling origami actuator with embedded capacitive sensors for proprioception, demonstrated via closed-loop position control and an open-loop three-actuator balance board capable of supporting a rider. The authors develop a kinematic model with a linear inverse relation $L = 0.22\xi + 10.4$ mm (with $\xi$ in degrees) and achieve high sensor sensitivity through finite element analysis-driven electrode optimization, achieving $R^2$-accurate length estimation without vision. The work provides a complete design-and-control workflow, including CAD generation, a dedicated fluidic control board, ROS nodes, and open-access code and CAD files to advance portable soft robots for everyday tasks.

Abstract

An underdeveloped capability in soft robotics is proprioceptive feedback control, where soft actuators can be sensed and controlled using only sensors on the robot's body. Additionally, soft actuators are often unable to support human-scale loads due to the extremely compliant materials in use. Developing both feedback control and the ability to actuate under large loads (e.g. 500 N) are key capacities required to move soft robotics into everyday applications. In this work, we independently demonstrate these key factors towards controlling and actuating human-scale loads: proprioceptive (embodied) feedback control of a soft, pneumatically-actuated origami robot; and actuation of these origami origami robots under a person's weight in an open-loop configuration. In both demonstrations, the actuators are controlled by internal fluidic pressure. Capacitive sensors patterned onto the robot provide position estimation and serve as input to a feedback controller. We demonstrate position control of a single actuator during stepped setpoints and sinusoidal trajectory following, with root mean square error (RMSE) below 4 mm. We also showcase the actuator's potential towards human-scale robotics as an "origami balance board" by joining three actuators into an open-loop controlled system with a platform that varies its height, roll, and pitch. This work contributes to the field of soft robotics by demonstrating closed-loop feedback position control without visual tracking as an input and lightweight, soft actuators that can support a person's weight. The project repository, including videos, CAD files, and ROS code, is available at https://parses-lab.github.io/kresling_control.

Figures (11)

• Figure 1: The origami actuator (a) overview and (b) under fluidic actuator. Photographs of (c) the low actuator mass, (d) actuators at different scales, and (e) a rider preparing to step onto the demonstration balance board.
• Figure 2: (left) The fluidic control board, further discussed in section 6.2 (right) A photograph of the system components including: Kresling actuators, the base plate, and pneumatic tubing entering the side of the base frame.
• Figure 3: A photograph of the flexible TPU filament used for 3D printing.
• Figure 4: The parameters of the Kresling structure (a) from the bottom face looking up and (b) in 3D
• Figure 5: Finite element electrode modeling. (a) The initial electrode shapes IC1-IC5 and OC5. The optimized sensitivities are presented in the top left corners. (b) The simulation results for electrode shapes OC1$-$OC5 and the triangle electrode. Inset:$\Delta C_{65}$.