Bulletin of the American Physical Society
APS March Meeting 2022
Volume 67, Number 3
Monday–Friday, March 14–18, 2022; Chicago
Session K03: Robophysics IFocus Recordings Available
|
Hide Abstracts |
Sponsoring Units: DBIO GSNP Chair: Dan Goldman, Georgia Tech Room: McCormick Place W-176A |
Tuesday, March 15, 2022 3:00PM - 3:36PM |
K03.00001: The Physics of Interactions in Soft Robotics Invited Speaker: Cecilia Laschi Soft robotics is a relatively young field of robotics that takes advantage of soft materials and compliant structures to improve robot interactions with their surrounding environment and ultimately provide them with enhanced abilities. It is largely grounded on bioinspiration and the soft robotics community is traditionally interdisciplinary. At the base of the use of such a high deformability is the idea of Embodied Intelligence, i.e. how the behaviour emerges from the physical interaction of the body with the environment. The physics of embodied intelligence is then crucial for its effective implementation in soft robotics. In addition to the typical need for modelling the robot deformations as given by actuators, soft robotics needs modelling the external interactions as well, for considering the action of external forces onto the robot deformation and overall movements, in robot control and design. Bioinspired soft robots find diverse applications, as they can access remote areas, confined spaces, even inside the human body, complex or collapsed structures, both in land and at sea. So, a variety of environments with diverse physical characteristics are in play. Marine applications give interesting challenges for bioinspired swimming, locomotion and manipulation underwater, involving fluid-robot interactions. Principles from octopuses and crabs enabled the development of efficient legged robots, walking underwater with emergent self-stabilizing gaits. Likewise, octopuses show unconventional, efficient, grasping and manipulation movements. Soft robotics is enabling robot abilities that were not possible before, like morphing, stiffening, growing, self-healing, evolving. They open up new scenarios, in the direction of more life-like robots. Seeing soft robots as dynamical systems in those environments allows to find emergent behaviours, effective and efficient, in an interdisciplinary dialogue. |
Tuesday, March 15, 2022 3:36PM - 3:48PM |
K03.00002: Worm omega turn modeling and its limbless robot implementation via geometric mechanics Tianyu Wang, Baxi Chong, Kelimar Diaz, Yuelin Deng, Ruijie Fu, Hang Lu, Howie Choset, Daniel I Goldman Reorientation plays a key role in limbless biological and robotic locomotion on different types of terrains. However, modeling the reorientation behaviors found in biological systems and implementing them on robotic platforms remain challenging. This work focuses on an omega shaped turning motion of the mm-long nematode worm C. elegans that allows large and rapid reorientation in natural environments (e.g., rotting fruit). By investigating body curvature dynamics with PCA analysis, we model omega turns as a superposition of two parametrized traveling waves: a forward wave, which drives forward motion, and an omega wave, which triggers turning. Using tools from geometric mechanics, we analyze the motion generated by the two-wave model with numerical simulations and determine the optimal coordination of parameters (e.g., the spatial frequencies) that maximizes the turning angle. Inspired by the C. elegans omega turning motion model, we develop controllers that enable limbless robots to exhibit omega shaped turning for reorientation. By geometric simulation and robophysical experiment, we demonstrate that the omega turn-inspired controller allows a limbless robot to produce effective and robust turning motion on challenging terrains, such as rigid lattices and granular media. |
Tuesday, March 15, 2022 3:48PM - 4:00PM |
K03.00003: Locomotion planning and control for discretely-soft bodied legged robots Hari Krishna Hari Prasad, Kaushik Jayaram The distributed compliance of soft-legged robots enables them to explore complex environments using novel gaits with high levels of safety and adaptability. However, this makes designing, controlling, and motion planning for such systems often challenging due to a large number of unactuated/underactuated degrees of freedom. To address this issue, current solutions often include constraining their motion to a few predefined modes by design and typically require complicated data-driven system identification and controller design techniques. In this work, we present two geometric motion planning frameworks for an autonomous, discretely soft, closed kinematic chain articulated, insect-sized quadrupedal robot with electroadhesive feet - CLARI (Compliant Legged Articulated Robotic Insect). Inspired by geometric mechanics-based gait generation for serpentine robots, we develop 'gait controllers' for in-plane, `shape-shifting locomotion of CLARI. The ability to deliberately and precisely control the robot-ground interaction allows us to implement sequential 'electroadhered' strides of leg-pairs rendering the locomotion problem kinematic with exact solutions for robot configurations, and global trajectories. We expect this approach to generalize to n-linked robots with m ($<$n) legs. |
Tuesday, March 15, 2022 4:00PM - 4:12PM |
K03.00004: Snake robot uses vertical bending with force feedback control to traverse large obstacles Qiyuan Fu, Chen Li Snake robots are promising in traversing 3-D terrains as well as snakes do. Despite many previous studies using lateral bending for propulsion, snake robots still struggled on 3-D terrains using these gaits even with compliance to adapt to height changes. Vertical bending can push against terrain height variation for propulsion (Date and Takita, 2007 IROS; Jurestovsky et al., 2021 JEB; Fu et al., 2021 SICB), which can potentially dramatically expand the source of propulsion to improve performance in 3-D environments. To understand how to effectively utilize vertical bending, we use a snake robot to test two hypotheses: (1) With large height variation, vertical bending alone can generate substantial propulsion; (2) With force feedback, propulsion can be better maintained. To test the first hypothesis, we controlled the robot to traverse a large obstacle under different backward loads. The robot generated propulsion 3 times its own frictional drag by propagating a fixed vertical bending that matched obstacle shape down the body. To test the second hypothesis, we challenged the robot with unknown obstacles and checked whether force feedback control improved performance. The robot indeed maintained better contact with terrain without reducing active pushing when using force feedback. |
Tuesday, March 15, 2022 4:12PM - 4:24PM |
K03.00005: A Modular Approach to Discrete Differential Geometry-based Simulation of Soft Robots Mohammad Khalid Jawed, Yayun Du, Mrunmayi Mungekar Owing to their resemblance to natural organisms in terms of structural compliance and reversibility, soft robots can navigate through unstructured environments via interaction with the environment. A bottleneck for real-world application of such robots is the lack of accurate and efficient predictive modeling tools. Geometric and material nonlinearity of the structure, and the complex interaction with the surrounding medium, pose challenges to comprehensively understanding and predicting the movement of the robot. As a result, soft robots are almost invariably designed and controlled through a cumbersome trial and error process. The most typical soft robots is comprised of a number of slender structures. Here, we present a numerical simulator for articulated soft robots inspired by a discrete differential geometry-based computational framework. Simulating several robotic testbeds, it runs faster than real-time on a single thread of a modern desktop processor. The simulator features an implicit approach for accounting for material elasticity, geometric nonlinearity and the motor actuation, and can easily incorporate external forces such as gravity, hydrodynamics, and magnetic forces. We explore the physics of locomotion in granular media and viscous fluids using our simulation tool and an articulated robot testbed with multiple elastic tails. Our experiments and simulations show reasonable quantitative agreement, implying that this discrete geometric approach could be used as a computational framework for predictive simulations of soft robot design and control. |
Tuesday, March 15, 2022 4:24PM - 4:36PM |
K03.00006: Coordinating tiny limbs and long bodies: geometric mechanics of diverse undulatory lizard locomotion Baxi Chong, Tianyu Wang, Eva Erickson, Philip J Bergmann, Daniel I Goldman Although typically possessing four limbs and short bodies, lizards have evolved diverse body plans including elongate trunks with tiny limbs. These elongate morphologies are hypothesized to aid locomotion in cluttered and fossorial environments. However, mechanisms of propulsion in such forms – e.g. the use of body and/or limbs to interact with the substrate – and potential body/limb coordination remain unstudied. Here, we use biological experiments, a geometric theory of locomotion, and robophysical models to comparatively investigate body-limb coordination in a diverse sample of lizard morphologies. Locomotor field studies in short limbed, elongate lizards (Brachymeles and Lerista) and laboratory studies of fully limbed lizards (U. scoparia and S. olivaceus) and a snake (C. occipitalis) reveal that body wave dynamics can be described by a combination of standing and traveling waves; the ratio of the amplitudes of these components is inversely related to the degree of limb reduction and body elongation. The geometric theory helps explain our observations, predicting that the advantage of traveling wave body undulations (compared with a standing wave) emerges when the dominant thrust generation mechanism arises from the body rather than the limbs. We test our hypothesis in biological experiments by inducing use of traveling waves in stereotyped lizards via modulating the ground penetration resistance. Study of a limbed/undulatory robophysical model demonstrates that a traveling wave is beneficial when thrust is generated by body-environment interaction. Our models could be valuable in understanding functional constraints on the evolutionary process of elongation and limb reduction in lizards, as well as advancing robot designs. |
Tuesday, March 15, 2022 4:36PM - 4:48PM |
K03.00007: Bio-inspired sensing and control of soft slithering bodies in 3D environments Xiaotian Zhang, Tixian Wang, Andrew Dou, Prashant Mehta, Mattia Gazzola We present a bio-inspired neural architecture for learning closed-loop controls in slithering soft bodies, realized through coupled oscillators and reinforcement learning. We deploy this architecture in a soft snake model and demonstrate how the snake can filter and leverage noisy sensory measurements to navigate realistic 3D environments. |
Tuesday, March 15, 2022 4:48PM - 5:00PM |
K03.00008: Enhancing Manueverability via Gait Design Siming Deng, Ross L Hatton, Noah J Cowan Locomotion gait design typically focuses on optimizing for various notions of efficiency, such as cost of transport or speed. Equally important is the ability to modulate a gait (whether optimal or not) to steer the system. In drag-dominated locomotion, geometric mechanics provides an elegant and practical framework for both goals---gait design and gait modulation. This framework gives tools for approximating the net displacement of robotic systems over cyclic gaits and optimizing for the most efficient gaits according to a user-specified cost. In this work, we propose both local and global gait morphing algorithms for modulating a nominal gait to provide efficient, single-parameter steering control. Using a simplified swimmer, we numerically compare the two approaches and show that for modest turns, the local approach, while suboptimal, nevertheless proves effective for steering control. A potential advantage of the local approach is that it can be readily applied to soft robots or other systems where local approximations to the constraint curvature can be garnered from data, but for which obtaining an exact global model is infeasible. This work was also submitted to ICRA 2022 (in review). |
Tuesday, March 15, 2022 5:00PM - 5:12PM |
K03.00009: Granulobot: From granular matter to self-assembling and reconfigurable robotics Baudouin Saintyves, Matthew Spenko, Heinrich M Jaeger We introduce Granulobot, a soft robotic platform that uses the rich properties of granular matter to propose new resilient capabilities. It consists of many identical motorized units that embed a Wifi microcontroller and sensors. Each robot is a few centimeters in diameter and uses passive magnets to engage and disengage with others. By applying torques, individual robots can move in space, self-assemble, and reconfigure collectively. Such minimal principle encompasses a scalable, decentralized robotic system, that can change its morphology and its mechanical behavior in real-time. In a vertical configuration, the granular organization provides structural strength against gravity, and changing the locally applied torques can transform the behavior from rigid to plastic and compliant. The control of such complex system, with many degrees of freedom, challenges common strategies used in robotics. Through examples, we explore control approaches based on physical properties and decentralized feedback. With Granulobot, we expend upon the complexity of many bodies systems to develop a form of “morphological” computation. This allows the design of resilient and autonomous systems, with promising perspectives in extreme conditions applications (space exploration and maintenance, catastrophe management, etc.). |
Tuesday, March 15, 2022 5:12PM - 5:24PM |
K03.00010: Manipulating propulsive forces with mulitple fins: Understanding how the relative location, phase relationship, compliance and flapping frequency of fins affect the swimming forces and associated flows Anthony Mignano Fish produce swimming and maneuvering forces through the coordinated motions of their body and fins. To achieve a locomotory task, fish will typically use multiple fins cooperatively with some fins interacting with the wakes shed by the body and other fins. Since the fin locations, kinematics and mechanical properties vary for different fish species, and with multi-finned robots being more frequently found in research, a better understanding of the complex hydrodynamic interactions among the fins and body that affect propulsive forces is needed. Several studies using biorobotic and numerical models were performed to understand how the relative location, phasing, compliance and flapping frequency of fins affected the produced forces and the associated fin-wake interactions. The results of these studies showed that each experimental factor had a distinct effect on the magnitude, shape and range of propulsive forces achievable by changing the fin kinematics; and the effect of each factor is coupled via the hydrodynamic interaction between fins and wakes shed by other fins. Understanding and capitalizing on the interdependency of fin phasing, spacing, compliance and flapping frequency is essential if engineers are to successfully design and operate multi-finned robotic systems. |
Tuesday, March 15, 2022 5:24PM - 5:36PM |
K03.00011: Mismatch of body undulation and limb waves enables robust centipede locomotion Juntao He, Tianyu Wang, Baxi Chong, Kelimar Diaz, Eva Erickson, Daniel I Goldman Centipedes that use retrograde (head-to-tail) limb stepping waves display body undulation when locomoting at high speed (Manton et al. 1952). Yet, little is known of how centipedes coordinate body undulation with stepping patterns to locomote. We studied body-limb coordination in centipedes via a robophysical model (L=70 cm, 8 leg pairs), programmed to generate retrograde waves for both the body and limbs. This robot achieved effective gaits when body-limb waves had equal spatial frequency (i.e.,nbdoy=nleg). However, the gait efficacy was sensitive to body-limb coordination if nbdoy=nleg=1. Specifically, changes (e.g., ±π/3) in the body-leg wave phase lag led to a decrease (e.g., 40%) in speed. With nbdoy=nleg=1 the average speed across all body-leg phase lags (0 to 2π) was 8.6±17.1 cm/cycle. In contrast, by increasing nleg while fixing nbdoy, the robot’s locomotive performance was robust to changes in the body-leg wave phase lags. With nbdoy=1 and nleg=1.6, the average speed across all body-leg phase lags was 15.4±7.0 cm/cycle. A similar mismatch in body-leg spatial frequency (nbody=1.6±0.2, nleg=2.2±0.2) was observed in the S. polymorpha (N=3, L=7.3±1.5cm, 19 leg pairs). This suggests robust centipede locomotion is achieved by distinct frequencies in the body and limb waves. |
Tuesday, March 15, 2022 5:36PM - 5:48PM |
K03.00012: Studying force generation and control in snakes using a sensorized robot Divya Ramesh, Qiyuan Fu, Chen Li Snakes can combine vertical and lateral body bending to adapt to and traverse irregular 3-D terrain with obstacles much larger than their body height with little slip and instability, an ability that most snake robots still lack. Observation of snakes adapting to and traversing vertical structures on flat surfaces suggests that this ability likely relies on contact force sensing and feedback control. To begin to understand this, we created a 12-segment robot with each segment instrumented with 3 sheet contact sensors capable of measuring normal contact forces at 30 Hz. We chose to use piezo-resistive sensors considering that they are cheap, light, flexible, and can be made to desired shape to fit outside and distributed on robot body. Each segment has a soft shell to distribute force evenly on sensor by conforming to terrain. The robot traversed a large obstacle by propagating down its body a predefined vertical bending shape that conformed to the obstacle. We observed large normal contact forces on the segments in contact with front side of obstacle, resulting in net forward propulsion to overcome frictional drag. Because the sensor's piezo-resistive material displays a creep behavior when loaded, we are working on a sensor model to accurately estimate dynamic contact force. |
Tuesday, March 15, 2022 5:48PM - 6:00PM |
K03.00013: Emergence of undulation gait based on embodied position controller and filter Longchuan Li, Shugen Ma, Isao T Tokuda Undulation is the most common gait generated by legless creatures, which enables their robust and efficient locomotion in various environments. Such advantages inspired the control design of many kinds of locomotion robots. Despite their technical details, most of them realize the undulation gait via tracking a predetermined trajectory called serpenoid wave, which is a group of sinusoidal functions with specified phase differences. This approach, however, sounds quite redundant in terms of sensing and control. Therefore, we come up with two research questions: 1. Whether the undulation gait should be predetermined, or it can naturally emerge? 2. Whether sinusoidal waveform is necessary to be encoded in the control signal to make the whole body an “S-shape”? Accordingly, theoretical analysis is performed via dynamic morphological computation. The results show that undulation emerges based on embodied position controller and filter, where binary actuation signals with clocks are required only. The findings not only discover locomotion mechanisms for significantly reducing the sensing and control requirement of generating artificial undulation gait, but also provide additional understandings for biological systems from the mechanical engineering point of view. |
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