Bulletin of the American Physical Society
APS March Meeting 2023
Volume 68, Number 3
Las Vegas, Nevada (March 5-10)
Virtual (March 20-22); Time Zone: Pacific Time
Session Q10: Robophysics IFocus
|
Hide Abstracts |
Sponsoring Units: DBIO Chair: Daniel Goldman, Georgia Tech Room: Room 202 |
Wednesday, March 8, 2023 3:00PM - 3:36PM |
Q10.00001: A hybrid Eulerian-Lagrangian simulation framework for heterogeneous soft structures operating in fluids Invited Speaker: Mattia Gazzola Structures encountered in biological and robotic domains are often constituted of slender elastic elements that are distributed, heterogeneous, and hierarchically organized. Their interaction with surrounding fluids is computationally challenging and expensive to resolve. Here, we mitigate these issues via a hybrid Eulerian-Lagrangian algorithm that combines Cosserat rod theory with vortex methods and immersed boundary techniques. The resulting elastohydrodynamic solver is tested against a battery of benchmarks, and further extended to the context of muscular hydrostats, multi-body contact, magnetic actuation, and immersed soft robot design. |
Wednesday, March 8, 2023 3:36PM - 3:48PM |
Q10.00002: Physical model inspired by energetic jumping nematodes Sunny Kumar, Victor M Ortega-Jimenez, Ishant Tiwari, Saad Bhamla Nematodes are a taxon of microscopic worms that are more abundant than all individual animals combined. The majority of nematodes use undulatory propulsion to swim or crawl across wet conditions. Impressively, entomopathogenic nematodes which parasitize insects, are unique among roundworms because they can jump. Although the kinematics of this jumping behavior has been identified nearly 60 years ago, the energetics of elastic energy storage and release have remained unclear. Here, we utilize soft robophysical elastic structures including polymeric-based elastic cylinders and fluid-filled balloons (shells) to explore how the hydrostatic skeleton, cuticle, and muscles act as non-linear springs to store energy in the loop formation of the worm body prior to jumping. We specifically focus on the role of kinks (sharp folds) that are formed when these elastic cylindrical structures are bent beyond its buckling limit. We show that kinks in these highly deformable bodies could serve multifunctional roles: acting as a “capacitor”, enabling slow energy build-up and fast release; creating a non-linear spring for low-force, yet high energy loading; and finally offering stability during loop formation. Our study sheds insight into both how organisms exploit elastic instabilities for ultrafast motions while offering design motifs for soft jumping robots. |
Wednesday, March 8, 2023 3:48PM - 4:00PM |
Q10.00003: Modeling Locomotion of a Hydrogel Crawler Siming Deng, Bibekananda Datta, Junning Liu, Thao (Vicky) Nguyen, Brian A Bittner, Noah J Cowan Soft smart materials used as actuators play an increasingly important role in the development of soft, biologically compatible locomotion systems. However, their compliant nature and distributed surface interactions make the systems highly complex. While soft body locomotion has been demonstrated at a variety of length scales, the modeling of such systems remains highly specific and ad-hoc. Data-driven geometric mechanics provides a practical framework for characterizing system dynamics for dissipative and underactuated systems. Here, we present a new application of data-driven modeling on a soft crawler made of thermo-responsive hydrogels, materials that swell and shrink as a function of temperature. Forward locomotion requires symmetry breaking, and most prior hydrogel crawlers rely on surface features to break symmetry; the design presented here uses the morphologically tuned, spatially asymmetric hydrogel swelling dynamics to induce locomotion, eliminating the need for specialized surface structures. For this specific system, we show that despite the complexity introduced by the soft body, its body shape can be characterized using a low dimensional shape subspace via straightforward dimensionality reduction (PCA). Based on finite element simulation data, we built and tested a data-driven model for the hydrogel locomotion behavior around its typical temperature cycles. The next step will be to test our locomotion modeling and gait design approach using physical hydrogel robots. |
Wednesday, March 8, 2023 4:00PM - 4:12PM Author not Attending |
Q10.00004: Kraken – A Multiphysics Soft Robot Simulation Platform Kevin Wandke, Y Z
|
Wednesday, March 8, 2023 4:12PM - 4:24PM |
Q10.00005: Deep Learning Force Manifolds from the Physical Simulation of Robotic Paper Folding Andrew Choi, Dezhong Tong, Demetri Terzopoulos, Jungseock Joo, Mohammad Khalid Jawed The focus on intelligent robotic manipulation of slender deformable objects has grown in popularity in recent years due to its numerous real world applications. A vast majority of prior works have used data-driven control policies that often rely purely on visual feedback. The lack of physical insight on such methods result in a critical lack of generality in terms of material, geometric, and/or environmental (e.g. friction) changes. In regards to this, we propose a novel control strategy for the difficult task of single manipulator paper folding with complete generality. Physically-accurate simulation and machine learning are combined to train fast and accurate deep neural networks, referred to as "force manifolds", capable of predicting the external forces of the paper given a grasp position. Scaling analysis is performed to obtain a problem formulation independent of material and geometric properties. Path planning is then used over the manifold to produce optimal robot folding trajectories. The high inference speed of our train models allow for real-time visual feedback resulting in closed-loop sensorimotor control. We demonstrate significant improvement over natural folding strategies for papers of various materials and shapes through extensive real world experiments. |
Wednesday, March 8, 2023 4:24PM - 4:36PM |
Q10.00006: Single-vertex origami for multimodal locomotion Davood Farhadi, Laura Pernigoni, David Melancon, Antonio M Grande, Katia Bertoldi The ancient art of origami provides an ideal platform for the design of reconfigurable systems, since a myriad of shapes can be achieved by actively folding thin sheets along pre-defined creases. Here, we focus on one of the simplest building blocks of origami-based materials - the rigid, degree-four vertex - and investigate its interactions with a flat surface. We show that by introducing asymmetry through the arrangement of the folds, we can realize asymmetric frictional forces that enable locomotion. We first investigate how the geometry of the degree-four vertex affects locomotion and then exploit these findings to realize an origami crawler capable of moving along arbitrary trajectories. |
Wednesday, March 8, 2023 4:36PM - 4:48PM |
Q10.00007: A Principal Bundle Perspective on Differential Flatness in Complex Robotic and Biological Systems Jake Welde, Matthew D Kvalheim, Vijay Kumar In recent decades, research has achieved deep insight into the locomotion of those robotic and biological systems whose evolution can be described in terms of a principal connection thanks to their natural symmetries. Meanwhile, the property of differential flatness has afforded another powerful method for the control of underactuated systems, but a tractable means of finding the requisite flat output for a given mechanical system has remained elusive. Recently, we have shown that the principal bundle structure induced by symmetry also furnishes a powerful tool for flat output discovery. Specifically, we give a sufficient condition for the direct construction of flat outputs from any section of the system's principal bundle that is orthogonal to a computable distribution. These flat outputs are the group variables of a trivialization of the bundle (a "choice of gauge" in the physics parlance), thus they are equivariant and typically global or almost-global. This perspective yields insight into long-standing open questions on the role of symmetry in differential flatness, while also facilitating flatness-based modeling of complex systems such as free-flying multibody robots and airborne insects, encompassing modes of locomotion that transcend the classical principal connection model. |
Wednesday, March 8, 2023 4:48PM - 5:00PM |
Q10.00008: Topology, dynamics, and control of an octopus muscular hydrostat Arman Tekinalp, Noel M Naughton, Seung-Hyun Kim, Udit Halder, Prashant Mehta, William Kier, Mattia Gazzola Muscular hydrostats, such as octopus arms or elephant trunks, lack bones entirely, endowing them with exceptional dexterity and reconfigurability. Key to their unmatched ability to control nearly infinite degrees of freedom is the architecture into which muscle fibers are weaved. Their arrangements is, effectively, the instantiation of sophisticated mechanical programs that mediate, and likely facilitate, the control and realization of complex, dynamic morphological reconfigurations. Here, by combining medical imaging, biomechanical data, and direct numerical simulations, we synthesize a 3D computational analog of an octopus arm, and begin to unravel this complexity. We show how arm motions can be understood in terms of storage, transport, and conversion of topological quantities, effected by basic muscle activation templates. These, in turn, can be composed into higher-level control strategies that, compounded by the arm's mechanical compliance, are demonstrated in a range of object manipulation and retrieval tasks. Overall, this work significantly advances modeling and simulation abilities in the space of heterogeneous and active structures while exposing design and algorithmic principles pertinent to muscular hydrostats, with implications in biology, robotics, and control. |
Wednesday, March 8, 2023 5:00PM - 5:12PM |
Q10.00009: Connecting biological design principles and optimal mechanics Jake E McGrath, José R Alvarado Biological systems often express optimal mechanics because they have navigated through some complicated evolutionary fitness landscape. Recent studies have highlighted that the nonlinear design principles of biological actuators (skeletal muscle) exhibit certain mechanical performance advantages – for example, improved energy economy, improved stability, and reduced information entropy in the control effort. We hypothesize that the nonlinear mechanical design principles of biological systems manifest optimal mechanics. However, because of the inherent complexity of the neuroskeletomuscular system, the underlying physical mechanisms that provide these well documented mechanical advantages remains poorly understood. Taking a reductionist approach, we create a robo-physical model system of animal legged locomotion to connect the underlying mechanical design principles to the observed mechanical performance advantages. We construct a two degree-of-freedom robotic leg constrained to jump vertically in one dimension. By implementing feedback control, we actuate the hopper with nonlinear, bioinspired force-velocity characteristics found in the Hill model of muscle. Furthermore, using impedance control in the hopper's ankle to mimic the function of tendons in a biological system, we study how stiffness and damping on impact provide passive mechanical stability. By studying a robo-physical model system of animal legged locomotion, we take a reductionist approach to uncover how the underlying nonlinear physical mechanisms in biology provide mechanical performance advantages. |
Wednesday, March 8, 2023 5:12PM - 5:24PM |
Q10.00010: Design and Closed-Loop Motion Planning of an Untethered Swimming Soft Robot Using 2D Discrete Elastic Rods-based Physics Engine Xiaonan Huang, Zachary Patterson, Andrew Sabelhaus, Carmel Majidi, Khalid Jawed, Weicheng Huang, Kiyn Chin, Zhijian Ren Despite tremendous progress in developing soft robots in recent years, existing systems lack the mobility, model-based control, and motion planning capabilities of their piecewise rigid counterparts. As in conventional robotic systems, the development of versatile locomotion of soft robots is aided by integrating hardware design and control with modeling tools that account for their unique mechanics and environmental interactions. Here, a framework for physics-based modeling, motion planning, and control of an untethered swimming soft robot is introduced. This framework enables co-design in the simulation of robot parameters and gaits to produce effective open-loop behaviors and enables closed-loop planning over motion primitives for feedback control of a soft swimmer. This pipeline uses a discrete elastic rods physics engine that discretizes the soft robot as many stretchable and bendable rods. On hardware, an untethered aquatic soft robot that performs frog-like rowing behaviors is engineered. Hardware validation verifies that the simulation has sufficient accuracy to find the best candidates for sets of parameters. The simulator is then used to generate a trajectory library of the robot's motion that is used in real-time closed-loop path following experiments on hardware. |
Wednesday, March 8, 2023 5:24PM - 5:36PM |
Q10.00011: Hierarchical control and learning of a foraging CyberOctopus Chia-Hsien Shih, Noel M Naughton, Udit Halder, Heng-Sheng Chang, Seung Hyun Kim, Rhanor Gillette, Prashant Mehta, Mattia Gazzola Inspired by the unique neurophysiology of the octopus, we propose a hierarchical framework that simplifies the coordination of multiple soft arms by decomposing control into high-level decision making, low-level motor activation, and local reflexive behaviors via sensory feedback. When evaluated in the illustrative problem of a model octopus foraging for food, this hierarchical decomposition results in significant improvements relative to end-to-end methods. Performance is achieved through a mixed-modes approach, whereby qualitatively different tasks are addressed via complementary control schemes. Here, model-free reinforcement learning is employed for high-level decision-making, while model-based energy shaping takes care of arm-level motor execution. To render the pairing computationally tenable, a novel neural-network energy shaping (NN-ES) controller is developed, achieving accurate motions with time-to-solutions 200 times faster than previous attempts. Our hierarchical framework is then successfully deployed in increasingly challenging foraging scenarios, including an arena littered with obstacles in 3D space, demonstrating the viability of our approach. |
Wednesday, March 8, 2023 5:36PM - 5:48PM |
Q10.00012: Mean field theory for larval Drosophila peristalsis Jane Loveless, Greg J Stephens With the rise and proliferation of low-cost imaging and automated tracking solutions, we can now measure the posture dynamics of individual animals at unprecedented resolution. However, theoretical approaches that would allow us to predict and explain our observations at this scale and resolution are relatively underdeveloped, and are often frustrated by the complexity and diversity of microscopic interactions that underlie animal behaviour. We describe a first-principles theoretical understanding of animal posture, using the Drosophila melanogaster larva (fruitfly maggot) as a model system. The larva locomotes via peristalsis: it propagates localised, longitudinal compression waves along its rod-like body. Larval posture during peristalsis can be represented via a scalar strain field, allowing us to probe the locomotor dynamics via non-equilibrium classical field theory. We construct a mean field theory (MFT) for the larva’s strain field via gradient expansion of the field’s equation of motion, and an assessment of symmetry, stability, and scaling properties of terms in this expansion. Our MFT admits soliton solutions in which damping and driving effects balance to produce a localised compression wave, similar to peristaltic waves observed in the real larva. Renormalisation group calculations suggests that this MFT is self-consistent in the presence of weak fluctuations. |
Wednesday, March 8, 2023 5:48PM - 6:00PM |
Q10.00013: Endowing Soft Robots with Fluidic Counting Sergio Picella Perception and remembering the number of events is a crucial feature for living systems. Such features fundamentally contribute to the autonomy and intelligence of beings in the whole Eukaryota. Counting in soft robotics is typically achieved by sensing mechanical and/or fluidic changes in the system through electric signals captured by dedicated sensors. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700