Search for: All records

Soft manipulators, renowned for their compliance and adaptability, hold great promise in their ability to engage safely and effectively with intricate environments and delicate objects. Nonetheless, controlling these soft systems presents distinctive hurdles owing to their nonlinear behavior and complicated dynamics. Learning‐based controllers for continuum soft manipulators offer a viable alternative to model‐based approaches that may struggle to account for uncertainties and variability in soft materials, limiting their effectiveness in real‐world scenarios. Learning‐based controllers can be trained through experience, exploiting various forward models that differ in physical assumptions, accuracy, and computational cost. In this article, the key features of popular forward models, including geometrical, pseudo‐rigid, continuum mechanical, or learned, are first summarized. Then, a unique characterization of learning‐based policies, emphasizing the impact of forward models on the control problem and how the state of the art evolves, is offered. This leads to the presented perspectives outlining current challenges and future research trends for machine‐learning applications within soft robotics. 

You can print anything... or can you? 3D printing is an exciting new technology that promises to very quickly create anything people can design. Scientists who want to make soft robots, like Baymax from Big Hero 6^TM, are excited about 3D printers. Our team uses 3D printing to make molds to produce soft robots. Molding is like using a muffin tin to make cupcakes. But can you make anything with 3D printing or are there times when 3D-printed molds do not work? Just like a cupcake liner, 3D-printed molds leave ridges, like a Ruffles potato chip, in soft robots. These ridges are a weak point where cracks can form, causing the robot to pop like a balloon. To prevent this, we sometimes need to make our robots using very smooth molds made from metal. This article talks about when and how 3D printing is useful in making soft robots. 

Aggressive and accurate control of complex dynamical systems, such as soft robots, is especially challenging due to the difficulty of obtaining an accurate and tractable model for real-time control. Learned dynamic models are incredibly useful because they do not require derivation of an analytical model, they can represent complex, nonlinear behavior directly from data, and they can be evaluated quickly on graphics-processing units (GPUs). In this paper, we present an open-source Python library to further current research in model-based control of soft robot systems. Our library for Modeling of Learned Dynamics (MoLDy), is designed to generate learned forward models of complex systems through a data-driven approach to hyperparameter optimization and learned model training. Included in the MoLDy library, we present an open-source version of NEMPC (Nonlinear Evolutionary Model Predictive Control), a previously published control algorithm validated on soft robots. We demonstrate the ability of MoLDy and NEMPC to accurately perform modelbased control on a physical pneumatic continuum joint. We also present a benchmarking study on the effect of the loss metric used in model training on control performance. The results of this paper serve to guide other researchers in creating learned dynamic models of novel systems and using them in closed-loop control tasks. 

This paper describes a series of endurance and material property tests conducted on a pneumatic, fabric-reinforced inflatable soft actuator made of Dragon Skin 30 silicone, which exhibited performance variations during operation. It is important to understand the level of variation over time and how it affects the motions of the soft actuators. The tests were designed to investigate the repeatability and durability of the actuator by measuring changes in its trajectories after long working periods, determining its failure pressure, and examining its elasticity through tensile tests. The experiments were performed on multiple soft actuators, and the results show pertinent information about the variation in their motion and how it relates to the material behavior of the silicone. This information enhances our understanding of the real-world behavior of silicone soft actuators and enables us to better control their performance in our applications. 

Optical waveguide deformation sensors are created for less than 15 US Dollars each and evaluated for their usefulness in detecting the severity of wrinkles in a thin-walled soft robot. This severity is quantified by the bend angle produced in the robot. The sensors are integrated into the skin of the robot and tests are performed to determine their usefulness. The sensors prove to be able to accurately track the bend angle of the robotic arm as a wrinkle is induced in a sudden load drop test, a sudden pressure loss test, an incremented load test, and an incremented pressure test. The drop test, specifically, tracked bend angle through many rapid undulations. 

This paper proposes a new method to measure the pose and localize the contacts with the surrounding environment for an inflatable soft robot by using optical sensors (photocells), inertial measurement units (IMUs), and a pressure sensor. These affordable sensors reside entirely aboard the robot and will be effective in environments where external sensors, such as motion capture, are not feasible to use. The entire bore of the robot is used as a waveguide to transfer the light. When the robot is working, the photocell signals vary with the current shape of the robot and the IMUs measure the orientation of its tip. Analytical functions are developed to relate the photocell signals and the robot pose. Since the soft robot is deformable, the occurrence of contact at any location on its body will modify the sensor signals. This simple measurement approach generates enough information to allow contact events to be detected and classified with high precision using a machine learning algorithm.