Session B25: From Responsive Matter to Actuated Structures

Decentralized reinforced learning of emergent behavior in robotic matter

Soft robots have the potential to be more robust, adaptable, and safer for human interaction than traditional rigid robots. State-of-the-art developments push these soft robotic systems towards applications such as rehabilitation and diagnostic devices, exoskeletons for gait assistance, and grippers that can handle delicate objects. However, despite these exciting developments, their inherent non-linear response limits the number of actuators that can be accurately controlled simultaneously, especially in complex or unknown environments. To enable modularly scalable and autonomous soft robots we have developed a new type of soft robot that is assembled from identical 1D building blocks with embedded pneumatic actuation, position sensing and computation. In this robotic system, motility might emerge from local interactions, rather than from a central brain. Here we shows that we are able to implement decentralized learning in this system. Using a stochastic optimization approach, each building block individually adjusts its actuation phase continuously, in order to find the fastest way to move in a predefined direction. We show that even for larger number of modules, this robotic system is still capable of learning. As a result, the system is robust to damage, as it will adjust its behaviour accordingly.   

Collective behavior of BOBbots, a robotic active matter system

We introduce BOBbots (Behaving, Organizing, Buzzing Robots, named to honor the late Bob Behringer), a custom low-cost active matter system composed of disk-shaped “granular” robots. The robots are equipped with controllable vibration that can react to light and stress. The vibration couples to two bristles on the base of the robot which generates translation and rotation. The robots can attract each other via loosely constrained magnets in cavities on a robot’s circumference. We first investigate the clustering of the collective which results from a competition between the attraction and driving. In experiments and discrete element method simulations, increasing magnet strength leads to a rapid increase in steady-state cluster size which saturates for sufficient attraction. The dynamics of this nonequilibrium system resemble those in an equilibrium lattice model. An algorithm by which robots decrease drive upon increased stress increases clustering rate at fixed magnet strength. Formation of clusters enables the collective to perform a task no individual can: transport of a free obstacle.   

Smarticle glider: Locomoting, spontaneous excitations in a shape changing active matter system

Smarticles or smart active particles are 3 link robots which change their shape periodically through the actuation of the arms. Limited individually in their ability to translate or rotate in the current experimental setting, they exhibit interesting collective behavior (directed locomotion) as a result of stochastic mechanical interactions (W, Savoie et al., Science Robotics 2019). The glider, which is a two smarticle bound excitation, is the fundamental building block for these phenomena. It has a lifetime of about 90 periods and moves about 4 body lengths before dissociating. We seek to understand the mechanistic principle behind the creation of these states and their ensuing dynamics with an iterative map. Simulations reveal that for these excitations to locomote, there must be fluctuations in the coefficient of friction. Since the smarticles only interact with each other via collisions, gliders are essential to keep a cloud of smarticles connected and hence harness their collective behavior.
Reference:
A robot made of robots: Emergent transport and control of a smarticle ensemble, W. Savoie et al., Science Robotics 2019.   

Photothermal actuation of a fluidic soft muscle

Despite innovations in materials and fabrication schemes that have enabled impressive soft robotic structures and devices, most soft robots remain tethered to cumbersome power sources (e.g., compressed gas) and/or require unwieldy valving. Developing soft robots capable of untethered actuation and remote control requires new methods of actuation. Compared to conventional power and control, visible light offers several advantages—including wireless remote control, spatial (~200 nm) and temporal (~ms) precision, and tunability (e.g., wavelength, intensity). Here, we exploit a novel class of photoswitches, donor-acceptor Stenhouse adducts (DASAs), for the remote actuation of soft muscles. Guided by studies of DASA’s unique photoswitching and photothermal properties, we achieve the remote actuation of a soft muscle via controlled photothermal phase changes without valves or internal controls. Further, we demonstrate the force output of this muscle and its promise for actuation in multi-scale soft robots.   

Programmable Photothermal Actuation using Novel Negative Photochromic Donor-Acceptor Stenhouse Adduct (DASA) Polymers

Light-driven actuation has significant advantages including untethered operation, precise spatiotemporal activation, and the ability to operate in complex surroundings without significant modification. Here, we use a new class of molecular photoswitches, called donor-acceptor Stenhouse adducts (DASAs), to generate a novel but conceptually simple photo-responsive polymer bilayer actuator, capable of repeatedly lifting a load against the force of gravity. We will present a synthesis pathway for chemically attaching DASA conjugates to poly(hexyl methacrylate) through norbornadiene click chemistry, and will demonstrate actuator performance. Importantly, we can leverage the different time scales of photothermal and photochemical responses of DASAs to achieve dynamic material control, including the ability to switch on and off actuation. Our results highlight the promising benefits of high molar absorptivity, negative photochromism, and visible light absorption of DASAs for actuation.
  

The realm of magnetorheological elastomers: experiments, theory and instabilities

In this talk, I will present a broad overview of our work on magnetorheological elastomers (MREs) including experiments, theory and numerical implementation. MREs are composite materials comprising metallic soft (e.g. iron) or hard (e.g. NdFeB) magnetic micron-sized particles. The latter have been called h-MREs. Out of these materials one can devise a number of interesting meso- and macrostructures, slender or not that can lead to a number of functionalities such as surface patterning, negative or positive swelling, network activation, negative Poisson ratio, evolving anisotropic magnetic properties, hierarchical buckling and others. The main idea lies in the ability to drive the systems by using two external loads, i.e. mechanical and magnetic ones and in particular to harness the resulting instabilities that are natural in magnetic structures. This allows, in turn, to reach a critically stable state by applying one of the fields (e.g. mechanical) and trigger patterning and instabilities at low energetic cost with the other (e.g. magnetic). As an example, we will discuss in more detail a recently obtained pattern, called crinkling. This pattern appears in slender magnetic structures such as MRE films bonded to soft (non-)magnetic substrates that are subjected to a combination of mechanical pre-compression and transverse to the film magnetic loads. In particular, the initial wrinkles evolve into crinkles, a bifurcation mode that is defined by the accompanied curvature localization and significant shearing of the side faces of the wrinkled geometry. Interestingly, the presence of the magnetic field prohibits the formation of creases and folds. The presentation will close with recent results on h-MREs and their potential ability to lead to unconventional responses mechanically and magnetically.   

Understanding the kinetic arrest in deswelling microgels

Biofriendly microgels, like PNIPAM and dextran, find numerous applications as microscopic delivery agents for advanced therapies of cancer, diabetes, and bacterial infections. We use mesoscale computer simulations to model the deswelling kinetics of spherical microgels. Deswelling occurs as the result of gel-solvent phase separation, leading to the hydrogel network collapse and in certain conditions, to the formation of a dehydrated “skin layer” at the gel-solvent interphase. This dehydrated skin layer hinders solvent flow from the microgel interior which arrests the kinetics of the collapsing microgel. We examine the conditions at which the skin forms and probe how these conditions depend on the microgel size, porosity, and micromechanics. These findings can be useful for designing microgel based drug release systems with controlled release profiles, leading to more effective disease treatments.   

Tunable buckling strength of magnetically active elastomeric shells

It has long been recognized that the buckling of shell structures is highly sensitive to material or geometric imperfections, leading to observed critical loads that are significantly lower than classic predictions. In this class of problems, the knockdown factor is defined as the ratio between the experimentally measured critical load and the classic theoretical prediction. This knockdown is typically regarded as an intrinsic property of the structure since the type and distribution of defects are encoded into the shell during fabrication. Here, we demonstrate the ability to actively tune the knockdown factor of pressurized spherical shells. We fabricate our spherical shells with a magneto-rheological elastomer (MRE) using a coating technique. The shells are first magnetized and then loaded by pressure under a uniform magnetic field. We find that, by adjusting the strength and polarity of the field, the knockdown factor of the magnetically active shells can be increased or decreased up to a maximum change of 30%. As such, we can externally tune their intrinsic buckling strength, on-demand. An axisymmetric shell model is used to rationalize the experimental results on how the magnetic field interacts with the buckling of imperfect shells.   

Tailoring magneto-mechanical properties of NdFeB particle-filled soft elastomers

This work investigates the possibilities of tailoring the magneto-mechanical properties of hard (permanent) magnetic NdFeB particle-filled soft magnetorheological elastomers by proposing novel microstructures. Our numerical computations for the mechanical response of various regular and chiral unit cells with a permanently magnetized elliptic inclusion show distinct features depending on the aspect ratio and the orientation of the inclusion. It is observed that the mechanical deformations are selectively preferred at certain directions, which make the magnetic inclusions align themselves along their easy axes, reducing the net magnetic moment. In particular, for the microstructural orientation angle of 30-60 degrees, the simple shear mode is observed to facilitate easy magnetic alignments and thus becomes energetically favorable over the uniaxial tension/compression mode. Moreover, the overall (effective) magnetic coercivity of these composites are found to reduce considerably with a softer elastomeric matrix, which, in turn, facilitate easy particle alignments.   

Magnetically triggered thin-film instabilities for surface pattern control

The study under consideration is concerned with instabilities of a thin magnetic film attached to a soft (magnetic) substrate under general biaxial precompression and transverse magnetic fields leading to complex two-dimensional surface patterns. This study builds upon previous experimental work on one-dimensional surface patterns through magneto-elastic loading.
Our findings contribute a key ingredient for the design of haptic displays and more general applications of surface actuation. For the characterization of the instabilities of interest we conduct numerical studies using finite-element and Fourier discretizations for a wide range of in-plane prestresses and applied magnetic field. By virtue of these results we obtain stability maps which are an important input for subsequent numerical and experimental studies in the post-buckling regime.