Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Design and Control of Modular Soft-Rigid Hybrid Manipulators with Self-Contact

Zach J. Patterson, Emily Sologuren, Cosimo Della Santina, Daniela Rus

TL;DR

This work tackles the limited payload of soft robots by introducing soft-rigid hybrid manipulators (SRH) that combine soft continuum segments with rigid joints to achieve tunable stiffness and enhanced load-bearing. The authors develop a PCC-based dynamic model, a configuration-space PD+ controller with a self-contact compensation term ${F}_c$, and a Cartesian impedance controller, demonstrating stable low-level control on hardware. Self-contact between rigid plates enables discrete, large increases in stiffness, enabling rigid-like behavior when needed while maintaining softness elsewhere. Experimental results include stiffness modulation measurements, load-bearing demonstrations, obstacle-course tasks, and robust disturbance rejection, highlighting the SRH paradigm as a practical path to safe yet capable robotic manipulation with high payloads. The work provides modular design, modeling, and open-source software to advance soft-rigid hybrids for real-world applications.

Abstract

Soft robotics focuses on designing robots with highly deformable materials, allowing them to adapt and operate safely and reliably in unstructured and variable environments. While soft robots offer increased compliance over rigid body robots, their payloads are limited, and they consume significant energy when operating against gravity in terrestrial environments. To address the carrying capacity limitation, we introduce a novel class of soft-rigid hybrid robot manipulators (SRH) that incorporates both soft continuum modules and rigid joints in a serial configuration. The SRH manipulators can seamlessly transition between being compliant and delicate to rigid and strong, achieving this through dynamic shape modulation and employing self-contact among rigid components to effectively form solid structures. We discuss the design and fabrication of SRH robots, and present a class of novel control algorithms for SRH systems. We propose a configuration space PD+ shape controller and a Cartesian impedance controller, both of which are provably stable, endowing the soft robot with the necessary low-level capabilities. We validate the controllers on SRH hardware and demonstrate the robot performing several tasks. Our results highlight the potential for the soft-rigid hybrid paradigm to produce robots that are both physically safe and effective at task performance.

Design and Control of Modular Soft-Rigid Hybrid Manipulators with Self-Contact

TL;DR

This work tackles the limited payload of soft robots by introducing soft-rigid hybrid manipulators (SRH) that combine soft continuum segments with rigid joints to achieve tunable stiffness and enhanced load-bearing. The authors develop a PCC-based dynamic model, a configuration-space PD+ controller with a self-contact compensation term , and a Cartesian impedance controller, demonstrating stable low-level control on hardware. Self-contact between rigid plates enables discrete, large increases in stiffness, enabling rigid-like behavior when needed while maintaining softness elsewhere. Experimental results include stiffness modulation measurements, load-bearing demonstrations, obstacle-course tasks, and robust disturbance rejection, highlighting the SRH paradigm as a practical path to safe yet capable robotic manipulation with high payloads. The work provides modular design, modeling, and open-source software to advance soft-rigid hybrids for real-world applications.

Abstract

Soft robotics focuses on designing robots with highly deformable materials, allowing them to adapt and operate safely and reliably in unstructured and variable environments. While soft robots offer increased compliance over rigid body robots, their payloads are limited, and they consume significant energy when operating against gravity in terrestrial environments. To address the carrying capacity limitation, we introduce a novel class of soft-rigid hybrid robot manipulators (SRH) that incorporates both soft continuum modules and rigid joints in a serial configuration. The SRH manipulators can seamlessly transition between being compliant and delicate to rigid and strong, achieving this through dynamic shape modulation and employing self-contact among rigid components to effectively form solid structures. We discuss the design and fabrication of SRH robots, and present a class of novel control algorithms for SRH systems. We propose a configuration space PD+ shape controller and a Cartesian impedance controller, both of which are provably stable, endowing the soft robot with the necessary low-level capabilities. We validate the controllers on SRH hardware and demonstrate the robot performing several tasks. Our results highlight the potential for the soft-rigid hybrid paradigm to produce robots that are both physically safe and effective at task performance.
Paper Structure (6 sections, 1 theorem, 26 equations, 7 figures)

This paper contains 6 sections, 1 theorem, 26 equations, 7 figures.

Table of Contents

  1. Introduction
  2. Results
  3. Methods
  4. Supplementary Information
  5. Acknowledgements
  6. Stability proof for state controller

Key Result

Theorem 1

There exists a $K_{\mathrm{P}}$ and $K_{\mathrm{D}}$ such that the trajectories of the closed loop system (eq:cl) are uniformly globally bounded and converge to the set $B_b = \{ e \in {R}^n \: : \: || e(t)|| \leq b\}$.

Figures (7)

  • Figure 1: A modular soft-rigid hybrid arm that can operate as a series of rigid bodies or soft segments. The figure shows a control workflow that incorporates contact compensation based on forward kinematics calculations. (a) Each module is controlled by a set of three tendons (visible ones on the first module are highlighted in red). (b) When a module is uncompressed, the module acts like a soft segment, particularly under load. (c) When the module is completely compressed, it acts as a de facto rigid body.
  • Figure 2: Design overview of the soft-rigid manipulator: (a) Panel visualizing a soft-rigid module transitioning between its soft and rigid states. As the cables contract, the rigid plates are pushed closer together, increasing overall stiffness. (b) A labelled soft-rigid module that relies on three dynamixel motors, spools, and cables to achieve various levels of stiffness and shape. (c) A single 3D printed hexagon plate, featuring its struts for better foam adhesion and through-holes for cables. (d) A full body rendering of the soft-rigid arm, composed of four soft-rigid modules.
  • Figure 3: Characterization of Soft-Rigid Module: (a) Plot of the bending stiffness test results. Inset is a rendering of the test performed on a module of the soft-rigid arm. (b) Planar slice of the manipulator workspace (which is radially symmetric). (c) Demonstration of manipulator supporting a 300 gram mass while the modules are in a rigid state (left) and a flexible state (right).
  • Figure 4: Disturbance Rejection and Trajectory Tracking: Step response and disturbance rejection for (a) modules and (b) joints. The inset images are snapshots of the robot as it is perturbed from the set point. (c) Torques output by the controller. Trajectory Tracking for (d) modules and (e) joints. The inset images are snapshots during the trajectory. (f) Torques output by the controller.
  • Figure 5: Impedance Control: Depicts an experiment in which impedance controller (\ref{['eq:imp']}) renders a desired spring damper system at the end effector compared to desired trajectory. (a) Cartesian coordinate of the end effector. (b) Control inputs. (c) Module state trajectory. (d) Joint motor state trajectory. (e) Snapshots from the trial.
  • ...and 2 more figures

Theorems & Definitions (2)

  • Theorem 1
  • proof