Bulletin of the American Physical Society
APS March Meeting 2024
Monday–Friday, March 4–8, 2024; Minneapolis & Virtual
Session F38: Robophysics IIFocus
|
Hide Abstracts |
Sponsoring Units: DBIO DSOFT Chair: Nick Gravish, University of California, San Diego Room: 103D |
Tuesday, March 5, 2024 8:00AM - 8:36AM |
F38.00001: In Search of Emergent Problem Solving in Intelligent Active Matter Invited Speaker: Robert H Austin Active matter is a field where motile agents powered by energy dissipative mechanisms move out of thermal equilibrium and form novel, emergent patterns and dynamics, but typically there is no intelligence to the system, each agent simply behaves as an automaton. Typically the agents have no sensors or the ability to make decisions, their motions are entirely due to local physical laws. Concommitent with the lack of decision making ability, the agents cannot communicate with each other except physically via physical forces. We can call this dumb active matter. We are developing a form of intelligent active matter where the agents communicate with each other over extended distances, have memories and code running internally where their actions are based on previous experiences and their internal records of changing environments which both sustain them and are hostile. The agents are deliberately a mix of digital and analog computers, so we believe the collective dynamics cannot be easily simulated on a conventional digital computer. Perhaps by 5 March we will succeed in showing that such an ensemble of computationally complex, interacting agents can begin to show the problem solving abilities that wet life forms can exhibit. |
Tuesday, March 5, 2024 8:36AM - 8:48AM |
F38.00002: Reprogrammable Microscopic Robots Maya M Lassiter, Jungho Lee, Kyle Skelil, William H Reinhardt, Lucas C Hanson, Li Xu, Dennis Sylvester, David Blaauw, Marc Z Miskin Here we show programmable, autonomous microscopic robots. Each robot integrates sensors, memory, electrochemical actuators for locomotion, solar cells for power, and a microcomputer into a single sub-mm machine. All these parts are built massively in parallel with fully lithographic fabrication. Robots move by generating electrokinetic flows, controlled by the on-board electronics. A user can program the robot by sending instructions via optical signals to an onboard receiver, digitally defining the robot’s behavior without physical changes. When running, the robot uses sensor data and/or data stored in memory for decision making, enabling sense-think-act loops and autonomous behaviors. For instance, we successfully wrote and implemented programs for gradient climbing, sensor value reporting, and gait control. Broadly, these programmable machines lower the barrier to microscopic robotics, allowing users to define and reconfigure a wide range of behaviors with a general-purpose robot too small to see by eye. |
Tuesday, March 5, 2024 8:48AM - 9:00AM |
F38.00003: Material Disturbance during Collective Construction with Soft Matter Joonha Hwang, Laura K Treers, Daniel Soto, Michael D Goodisman, Daniel I Goldman In collective construction, agents like ants and termites transport matter in their environment to create structures far larger than an individual. Instead of the deterministic approach that humans and other robot swarms employ, construction by invertebrate collectives is subject to uncertainty due to challenges in transport and cohesion of soft materials, agent interactions during collisions and crowding, and material disturbances from external and internal perturbations. Using a robophysical swarm, we seek to discover principles by which collectives can mitigate such challenges and form reliable, repeatable structures composed of soft materials. We use a minimalist two-armed, two-wheeled locomotor (24 cm long) equipped with force sensitive "jaws" to detect and transport cohesive materials (staples) from one side of a dark arena (1.8 x 1.2 m) where colored LEDs. In a cycle, a robot excavates then attempts to find the largest existing pile and place its excavated material. In early experiments, a single agent identified the primary deposition pile in 21 / 28 cycles, depositing 252 g of material per hour on average, leading to an emergent mound. In contrast, two agents using this algorithm deposited at an existing pile in only 8 / 24 cycles, depositing 142 g per hour per robot on average; importantly, this deposition was scattered in the arena. The decrease in excavation and deposition efficacy was a result of robot interference with each other and with deposited material (e.g., inadvertent scattering). |
Tuesday, March 5, 2024 9:00AM - 9:12AM |
F38.00004: Mind in Vitro: a complementary approach to physical intelligence Mattia Gazzola From the perspective of synthesizing fundamental principles of biological processing and control, robophysics approaches that emphasize the role of the embodiment have been successful in bridging animal science and engineering, revealing, for example, simple yet powerful strategies for gait organization, adaptation and locomotion, from snakes to fish. However, these creatures do possess a neural infrastructure that supports, synergizes and compounds physical intelligence. In order to aid dissecting principles of neural computations, pertinent to locomotion and beyond, here we discuss potential opportunities afforded by a 'Mind in Vitro' approach, whereby in vitro living processing architectures may be realized and operated by spatially distributing and connecting neural populations of various nature onto input/output/actuation substrates. |
Tuesday, March 5, 2024 9:12AM - 9:24AM |
F38.00005: The median statistics and velocity dependence of multi-contact dry friction systems Nick G Gravish, Rohan Shah The ground-based locomotion of animals and robots typically occurs through sequences of multi-contact interaction between body/limbs and the ground. Over dry, rigid surfaces the interaction force of a contact is governed by Coulomb friction, which is independent of the contact speed. However, recent experiments and simulation have revealed that multi-legged animals and robots exhibit dynamics that are consistent with speed-dependent interaction forces when walking [1], [2]. In this work we provide a simple demonstration of how this speed dependence arises in multi-contact frictional systems with heterogeneous contact velocities. We first present experiments from a multi-contact frictional carousel: a system of ten wheels that all rotate at varied speeds and collectively drive the rigid rotation of a carousel. We find that over more than 700 experiments in varying wheel speeds the net speed of the system is determined by the median of the wheel velocities. A simple theoretical model of a multi-contact system with speed-independent forces is able to reproduce this result from first principles. Furthermore, by examining the cumulative distribution function of instantaneous contacts we are able to derive an effective viscosity for multi-contact friction related to the gradient of instantaneous contact velocities. We study this effective viscosity in perturbation experiments of our multi-contact frictional carousel. Ultimately these experimental and theoretical results provide insight into how speed-dependent dynamics arise from individual speed-independent forces of multi-contact locomotion. |
Tuesday, March 5, 2024 9:24AM - 9:36AM |
F38.00006: The chaotic milling behaviors of interacting swarms after collision Sayomi Kamimoto, Jason M Hindes, Ira B Schwartz We consider the problem of characterizing the dynamics of interacting swarms after they collide and form a stationary center of mass. Modeling efforts have shown that the collision of near head-on interacting swarms can produce a variety of post-collision dynamics including coherent milling, coherent flocking, and scattering behaviors. In particular, recent analysis of the transient dynamics of two colliding swarms has revealed the existence of a critical transition whereby the collision results in a combined milling state about a stationary center of mass. In the present work, we show that the collision dynamics of two swarms that form a milling state transitions from periodic to chaotic motion as a function of the repulsive force strength and its length scale. We used two existing methods as well as one new technique: Karhunen–Loeve decomposition to show the effective modal dimension chaos lives in, the 0-1 test to identify chaos, and then constrained correlation embedding to show how each swarm is embedded in the other when both swarms combine to form a single milling state after collision. We expect our analysis to impact new swarm experiments which examine the interaction of multiple swarms. |
Tuesday, March 5, 2024 9:36AM - 9:48AM |
F38.00007: Excluded volume of polytopes and the geometry of inaccessible configurations Trevor F Teague, Sharon C Glotzer Determining the region of space that one object can't access due to the presence of another is a ubiquitous problem in physics and engineering, finding applications ranging from describing the ordering transitions of liquid crystals to performing path planning in robotics. Of particular interest in many domains is not only the region of inaccessible space but also its volume. For example, our understanding of emergent entropic transitions such as the crystallization of hard particles and the depletion interaction relies heavily on understanding how much space particles can and cannot occupy. Using the relevant theoretical tools from convex geometry we describe the excluded region and excluded volume concepts through the elementary yet practical perspective of convex polytopes. Formulas are derived for the excluded volume between two polytopes at fixed relative orientation, and the topology of the excluded region is explained using both mathematical arguments and the implementation of constructive algorithms. We utilize this theory to efficiently calculate the excluded volume of two convex polygons in 2D and polyhedra in 3D and describe its variation under translation, rotation, inversion, and dilation of one of the bodies. The theory developed in this work facilitates the investigation of the phase behavior of complex materials in which steric interactions play a dominating role. |
Tuesday, March 5, 2024 9:48AM - 10:00AM |
F38.00008: Stigmeric Coverage with a Robotic Collective for Surface Micro-patterning Annalisa T Taylor, Malachi Landis, Yaoke Wang, Ping Guo, Todd D Murphey Micro-structured surfaces enable beneficial properties including drag reduction and hydrophobicity. However, a lack of affordable and efficient manufacturing techniques prevent their widespread use. Currently, applying micro-scale divots to meter-scale surfaces requires expensive tooling to create evenly spaced indentations while supporting large workpieces. Here, we use mobile robots with credit-card-sized footprints for surface coverage to achieve high-fidelity patterns from detailed target images, patterning according to feature density instead of traditional feature specifications that require precision. In this work, we parallelize surface coverage with multiple decentralized robotic agents that indirectly communicate through stigmergy, a principle of coordination through altering and detecting signatures left by other agents in an environment. By sensing local feature density and comparing with a desired density function over the surface, robots can determine whether to prioritize texturing their immediate environment or cover area elsewhere. With this method, robots can collaborate to recreate detailed target images based on local information and indirect communication, providing an alternative manufacturing paradigm for surface texturing that is flexible and adaptable. |
Tuesday, March 5, 2024 10:00AM - 10:12AM |
F38.00009: Pulse coupled oscillators for micromachines Milad Taghavi, Wei Wang, Itai Cohen, Alyssa Apsel Microrobots are becoming increasingly functional and are poised to significantly impact various technical domains, including targeted drug delivery, precise surgical interventions, and environmental remediation. However, unlocking even greater potential lies in the ability of these microscopic robots to coordinate their actions and cooperate effectively. Achieving this coordination, however, presents a formidable challenge, primarily centered around the need for a scalable synchronization strategy—a crucial hurdle to overcome for enabling collective behaviors in multiple autonomous microscopic robots. |
Tuesday, March 5, 2024 10:12AM - 10:24AM |
F38.00010: Mechanically Transduced Soft Magnetic Rollers Joshua P Steimel, Nicholas Brown, Joseph Harrison, Alfredo Alexander-Katz Active soft autonomous systems integrated into biological systems is a new and emerging field with a myriad of potential applications. Here we present a novel soft active robotic system composed of functionalized ferromagnetic rollers. This system leverages the fundamental physical principle of friction at the microscopic level to autonomously control the amount of translational displacement of the ferromagnetic rollers. By functionalizing the ferromagnetic roller with biological ligands and a substrate coated with a corresponding binding ligand partner the friction coefficient is controlled by the strength and density of such biological binding interaction. The coefficient of friction will determine the amount of translational displacement. The ferromagnetic rollers are made active by actuation of an externally applied rotating magnetic and the rollers will proceed to translate as the rotational motion is converted to translational displacement due to friction between the ferromagnetic rollers and the substrate. The larger the coefficient of friction the larger the translational displacement and in this active soft robotic system the friction scales with the strength and density of the biological binding interaction. Here, we demonstrate this soft robotic system by functionalizing the surface of substrate with ubiquitin binding ligands and the rollers were coated with ubiquitin. Ubiqutin is one of the most ubiquitous proteins encountered in biological systems and we demonstrate how the displacement of the roller increases as the strength of interaction between ubitquin and the substrate ligands increases. This new approach opens up novel avenues for these soft robotic rollers to autonomously navigate biological soft substrates decorated with a myriad of different biological ligands. |
Tuesday, March 5, 2024 10:24AM - 10:36AM |
F38.00011: Differential Drive Kinematics from a Microscopic Robot William H Reinhardt, Lucas C Hanson, Maya M Lassiter, Tarunyaa Sivakumar, Scott Shrager, Marc Z Miskin Wheeled differential drives are one of the most widely used approaches in robotics thanks to easily controllable kinematics. Here, we present a robot three times smaller than a period of print that obeys the same kinematics, allowing us to import well developed control laws to the microscale. Each 300 micron robot has two photovoltaic-powered electrokinetic engines that enable it to locomote at ~1 body length per second. Using a closed-loop optical setup to track robots and pattern laser light, we can modulate the power on each engine to turn the robot and obtain controlled locomotion. We find the relationship between engine power and robot dynamics is identical to differential drive systems. Further, we implement standard control laws to autonomously pilot devices and obtain coordinated motion over multiple agents. Finally, we discuss ongoing work to expand this kinematic framework to include robot-to-robot interactions and collective behavior. |
Tuesday, March 5, 2024 10:36AM - 10:48AM |
F38.00012: Modeling of elastically coupled robots locomoting on an elastic membrane James K Lewis, Hussain N Gynai, Shengkai Li, Gongjie Li, Daniel I Goldman Based on previous work (Li et al., PNAS, 2022) involving a single two-wheeled robot (diameter 10 cm) locomoting on a deformable spandex membrane (diameter 2.4 m) we seek to model the rich dynamics of a modified system composed of two disk-shaped robots connected by a linear spring. Experiments reveal that the elastically connected vehicles can be captured in a trajectory where one vehicle always remains closer to a central depression. These “tidally-locked” dynamics are influenced by the coupling of the vehicles to their environment, their coupling to each other via the spring, and the interaction of the deformations they create in the membrane. To gain insight into the dynamics, we develop a numerical model with several important features. First, the model incorporates each vehicle’s differential drive which allows the wheels to have different drive forces, but enforces the constraint that the sum of the wheels’ drive forces remains constant. The interaction between the wheels and the elastic membrane is modeled by using experimentally measured drag forces on a constant speed wheel moving across the membrane at varying attack angles. We couple the vehicles to the environment using a measured mapping from membrane radius to centripetal force. The force between the vehicles is modeled as linear in the distance between robots. We find that the trajectories created by this model qualitatively match those seen in experiment and provide insights into the mechanism governing robophysical tidal locking. |
Tuesday, March 5, 2024 10:48AM - 11:00AM |
F38.00013: 'Mind in Vitro' open hardware: From versatile, robust single-unit neural interfaces to robot-assisted mini-clusters Xiaotian Zhang, Zhi Dou, Seung Hyun Kim, Gaurav Upadhyay, Hongbo Yuan, Zhantao Song, Daniel Havert, Kimia Kazemi, Scott Lee, Sehong Kang, Onur Aydin, Taher Saif, Nancy Amato, Hyun Joon Kong, John Beggs, Mattia Gazzola Motivated by the unexplored potential of in vitro neural systems for computing, control, and sensing, and by the corresponding need of versatile, scalable interfaces for multimodal interaction, we present accurate, modular, fully customizable and portable recording/stimulation solutions that can be easily fabricated, robustly operated, and broadly disseminated. Central to our approach is a single-unit neural interface that combines reconfigurable platforms, open-source electronics, and custom Micro-Electrode Arrays (MEAs). Designed with built-in versatility, the utility and robustness of our device is demonstrated across multiple neural culture types and applications. Finally, by coupling this single-unit apparatus with a robotic system, we demonstrate a mini neural computing cluster that allows automated and remote operation of tens of neural samples. Overall, our technology provides open-source and adaptable solutions to in vitro electrophysiology, paving the way to a novel neural computing paradigm. |
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