Session Y14: Robophysics: Robotics Meets Physics V: Actuation, Control, Design

Live

A fundamental problem with robotics at the microscale, is how to design robots that can be printed in 2D via lithography and yet are able to locomote and adopt arbitrary shapes. Auxetic mechanical metamaterials comprised of rigid panels that can locally splay, are an interesting solution to this problem because they can yield reconfigurable curved surfaces and generate different locomotion gaits for robotics applications. Here, we show that such electrically actuated auxetic metamaterials can be utilized to design micro-scale robots. The expansions and contractions in our devices are achieved by splaying neighboring panels. The actuation of the hinges is controlled by applying voltage to a nm thin surface electrochemical actuator. We modeled the target shapes using an inverse design approach in which the shapes are iterated towards target shapes by selecting optimal actuations of the splay hinges. We then show experimentally that we are able to generate a variety of 3D shapes by actuating a subset of the hinges. By integrating photovoltaics and timing circuits, we are working towards manufacturing untethered metamaterial-based micro-scale robots.   

Live

Sub-centimeter robotic systems have traditionally used piezoelectric bending actuators to actuate their various degrees of freedom due to their high-bandwidth and ease of control. However, they require high-voltage power electronics to operate, complex transmission mechanisms to amplify motion and thus, have a reduced force output limiting their ready implementation for actuating high-dimensional autonomous systems such as the limb of a legged robot to replicate animal-like dexterity. Voice-coil architecture based electromagnetic motors serve as attractive alternatives. Our origami-inspired, unidirectional, multi-scale fabrication technique leverages carbon fiber laminate and flexure based Sarrus linkage to function as both the housing and transmission unit, and when integrated into a single robot leg mechanism displays long-stroke high-force output and low-voltage operation. Preliminary modeling and experimental results show that our prototype (1.4cm in dimension), using a 3.175mm cubical N52 magnet produces about 50 times its body weight as static Lorenz force on a 41AWG, 46-turn, 4.75 mm diameter tightly wound coil with a constant input current of 0.1 A. Using impedance control, we propose to demonstrate various single-legged high-frequency hopping behaviors.   

Live

This study investigates a fundamental trade-off between a robot's state information usage and optimal decision making. Typically, robot state estimators and control systems are designed to produce and consume as much state information as possible — yielding a tight coupling between sensing and control.

This work identifies that it is often advantageous for the robot to minimize information usage to limit the effect of sensing error on control performance.We formalize the robot's control objective as a rate-distortion problem, and demonstrate that the optimal controller is given by a Gibbs measure. This structure allows us to quantify the performance of the controller when composed with a bounded-error state estimator.

Finally, we optimize the robot's information usage for a fixed level of estimation error and see that higher error levels produce lower information controllers.We conclude the study with corroborating simulations. Scenarios include hopping robots, ball catching, and object manipulation. Each example, demonstrates that low-information controllers perform robustly under estimation uncertainty.   

Live

There is a growing effort to understand decentralized control mechanisms, particularly in application to robotic systems with distributed sensors and actuators. Sea stars, being equipped with hundreds of tube feet, are an ideal model system for studying decentralized sensing and actuation. The activity of the tube feet is orchestrated by a nerve net that is distributed throughout the body; there is no central brain. We developed mathematical models of the biomechanics of the tube feet and the sea star body. We then formulated hierarchical control laws that capture salient features of the sea star nervous system. Namely, at the component level, the individual tube feet follow a state-dependent feedback controller. At the system level, a directionality command is communicated to all tube feet. We studied the locomotion gaits afforded by this control model. We find that these minimally-coupled tube feet coordinate to generate robust forward locomotion on different terrains. Our model also predicts different gait transitions consistent with our experiments performed on Protoreaster nodosus. These findings offer a new paradigm for walking using soft actuators, with potential applications to autonomous robotic systems.   

Live

We present a novel robotic manipulation capability for transporting an object on the ground in a dynamic and nonprehensile manner. The object is manipulated to rock from side to side repeatedly; in the meantime, the force of gravity enables the object to roll along a zigzag path that is eventually heading forward. We call it "rock-and-walk" object manipulation. Our work is motivated by an interesting question in archaeology, how the giant rock statues of Easter Island (known as "moai") were transported several hundred years ago, and a recent demonstration done by archaeologists that it is possible to "walk" the statue by repeated rocking. We examine the kinematics, statics, and dynamics of the rocking motion to understand how the state of the object evolves and how to control the robot to connect individual rocking motions into a stable gait of the object. Our rock-and-walk object transport technique is implemented in multiple robotic settings: 1) one robot arm with an end-effector that can cage the object, 2) two robot arms to interact with the object via cables, and 3) an aerial robot with the caging end-effector. A set of experiments demonstrate successful rock-and-walk object transport under a range of terrain conditions.   

Live

Self-righting when flipped over on the ground is strenuous for many terrestrial animals and robots. During self-righting, the discoid cockroach repeatedly opened and pushed its wings against the ground to attempt a somersault but rarely succeeded; instead, it often eventually rolled to the side to self-right. Its legs flailed in this process. Here, we studied whether simultaneous wing opening and leg flailing is beneficial. We developed a robot with two wings and a leg to emulate the animal’s strenuous self-righting. As wing opening and leg flailing amplitudes increased, self-righting probability increased. A potential energy landscape model to measure the potential energy barriers to self-right revealed the mechanism. Without leg flailing, the pitch kinetic energy from pushing wings was too small to overcome the high barrier to somersault. Without wing opening, the rolling kinetic energy fluctuation from the flailing leg was too small to overcome the low barrier to roll. However, when used together, wing opening reduced the roll barrier and enabled probabilistic crossing of the reduced roll barriers using kinetic energy fluctuation from leg flailing. Our study showed that animals and robots can modify their potential energy landscapes to better elicit locomotor transitions.   

Live

As the COVID-19 pandemic progresses, awareness of airborne virus-contamination and the importance of wearing a mask have increased dramatically. To minimize viral exposure for a long time, masks should be able to filter out sufficiently small droplets and have less breathing resistance. In this study, we proposed a 3D tortuous mask filter inspired by animal's nasal cavity. Based on Stokes number, our proposed filter with a 1 mm-width channel of a 1.8 tortuosity can capture small droplet particles of 5 µm or larger. Also, we demonstrated a system that monitors the user's breathing by a pressure sensor and controls an additional mask filter that opens and closes according to the level of breathing. When the pressure sensor catches the exceeding threshold, the regulator opens the additional filter to make human breathing easier. The performance of the mask filters and breathing regulator system was evaluated through a piston setup that imitated the human respiratory. Using this robotic system, the respiratory resistance and face-seal leakage rate were measured at the normal human breathing rate (15~100 LPM).   

Live

Soft robots have the potential to adapt their morphology, properties, and behavioral control policies toward different tasks or changing environments. This adaptive capability is often inspired by biological systems. For example, caterpillars display undulation and inchworm gaits but can rapidly curl themselves into a wheel and propel themselves away from predators. The armadillo can change from a walking gait on legs to a rolled-up ball as a defense mechanism. Octopus arms can access nearly infinite trajectories. During this talk, I will present recent work toward particulate and fibrous composites that address distributed sensing, variable stiffness properties, and variable trajectory motions inspired by these capabilities in animals. I will then contextualize the materials within three shape-changing robot platforms: robotic skins, robotic fabrics, and morphing limbs for amphibious locomotion.   

Live

Automated manipulation of multiple independently controlled Janus microrobots along with efficient and fast motion planning can potentially lead to a breakthrough in the formation of functional microstructures, both engineering and biological, in a scalable and reliable manner. In this talk, we present the first foundational step towards realizing this objective by developing a Q-learning-based controller which approximately optimizes the trajectories of the microrobots through rewards and discounts while avoiding obstacles, such as randomly dispersed micro-objects, or already formed microstructures. The self-learning microrobots can (1) recover a previously known propulsion policy, (2) identify a more effective policy, and (3) adapt the propulsion policy based on their interactions with the surrounding media. We show simulation experiments for independent motion and steering control for groups of one, two, and three microrobots, actuated using electromagnetic coils setup, to demonstrate the feasibility and effectiveness of our method.   

Live

The apple snail Pomacea canaliculata exhibits a unique feeding mechanism to collect food particles floating at the water-air interface: while under water, it positions part of its flexible foot parallel to the water surface and generates rhythmic undulations. These undulations trigger a flow near the free surface that brings the food particles towards the mouth. With a robotic system employing an actuation mechanism of the snail foot, we systematically unravel the fluid mechanics of this feeding mechanism. We observe that floating particles far away are sucked into the robotic snail. Through particle image velocimetry we quantify the velocity field around the actuating sheet for a range of capillary numbers. I will discuss how the size and speed of these undulations give rise to a pumping effect near the interface to drive the particle-laden fluid.   

Live

Marangoni stress plays an important role in many industrial applications where a surface tension gradient induces fluid flow. For instance, the Marangoni effect is incorporated in the cleaning process of silicon wafers in integrated circuit etching by applying a chemical gradient. The Marangoni effect can also be induced by a temperature gradient. In order to gain fundamental understanding about Marangoni flow in the absence of gravity, we analyze several different experimental videos of Marangoni experiments during the NASA STDCE-1 missions (performed on USML-Spacelab in 1995). We utilize multiple Particle Image Velocimetry and Particle Tracking Velocimetry methods to extract the flow field form NASA STDCE-1 videos and compare the experimental data to the numerical results of the flow-field. Finally, we discuss how our findings about temperature-driven Marangoni flow in the microgravity setting can improve cleaning processes in presence of gravity.