Keywords: Soft Robot Materials and Design, Modeling, Control, and Learning for Soft Robots, Biologically-Inspired Robots
Abstract: Aquatic creatures exhibit remarkable adaptations of their body to efficiently interact with the surrounding fluid. The tight coupling between their morphology, motion, and the environment are highly complex but serves as a valuable example when creating biomimetic structures in soft robotic swimmers. We focus on the use of asymmetry in structures to aid thrust generation and maneuverability. Designs of structures with asymmetric profiles are explored so that we can use morphology to `shape' the thrust generation. We propose combining simple simulation with automatic data-driven methods to explore their interactions with the fluid. The asymmetric structure with its co-optimized morphology and controller is able to produce 2.5 times the useful thrust compared to a baseline symmetric structure. Furthermore these asymmetric arms are validated on a robotic system capable of forward swimming motion while the same robot fitted with a plain feather is unable to move forward.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Applications, Soft Sensors and Actuators
Abstract: Electrically-actuated continuum soft robots based on Handed Shearing Auxetics (HSAs) promise rapid actuation capabilities while preserving structural compliance. However, the foundational models of these novel actuators required for precise control strategies are missing. This paper proposes two key components extending discrete Cosserat rod theory (DCM) to allow for modeling HSAs. First, we propose a mechanism for incorporating the auxetic trajectory into DCM dynamical simulations. We also propose an implementation of this extension as a plugin for the Elastica simulator. Second, we introduce a Selective Piecewise Constant Strain (SPCS) kinematic parameterization that can describe an HSA segment's shape with fewer configuration variables. We verify both theoretical contributions experimentally. The simulator is used to replicate experimental data of the mechanical characterization of HSA rods. For the second component, we attach motion capture markers at various points to a parallel HSA robot and find that the shape of the HSAs can be kinematically represented with an average accuracy of 0.3 mm for positions and 0.07 rad for orientations.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Applications, Optimization and Optimal Control
Abstract: Potential applications for large-scale soft robots include interacting with humans while carrying a heavy load, navigating in clutter, executing impact tasks like hammering a nail into a wall, and so much more. Because of their compliance and lack of fragile gear trains, soft robots are uniquely suited to these tasks. However, we expect that path planning may be more constrained by soft robot kinematics and dynamics than traditional rigid robots. Generating dynamically feasible trajectories for soft robots (especially large-scale soft robots with higher payloads) is critical to the success of low-level controllers tracking reference trajectories. This paper introduces an optimization method to generate task and joint space trajectories for soft robots that satisfy kinematic and dynamic constraints which are unique to large-scale soft robots. The method presented in this paper is an offline trajectory generator that is then fed to a low-level PID joint angle controller. We conduct two experiments to validate this method on a continuum pneumatic soft robot of length 1.19 meters in both simulation and on hardware. We show that this is a viable method of planning trajectories for soft robots with a reported median magnitude of error of 0.032 meters between the planned and actual end effector trajectories.

Keywords: Soft Sensors and Actuators, Soft Robot Applications, Soft Robot Materials and Design
Abstract: Soft robots are uniquely suited for applications in inspection, search and rescue, and exploration in confined and unstructured environments. Leveraging the benefits of these soft robots will increasingly require the integration of equally compliant electronic components for sensing, actuation, control, and computation to maintain mechanical conformability at the system level. In this work, we demonstrate the integration of modular electronic sensors into a robotic snake platform using stretchable interconnects. Aimed at applications in nuclear infrastructure inspection, included sensors include visual perception and on-board radiation spectroscopy. In addition to high mechanical compliance, the stretchable interconnect enables physical separation of analog sensors and digital electronics for use in radiation environments. We present details of stretchable electronics fabrication and integration, standalone validation of integrated sensors, and field test results from a simulated nuclear infrastructure inspection task.

Keywords: Soft Sensors and Actuators, Force and Tactile Sensing, Soft Robot Applications
Abstract: In this paper we present the design and experimental characterization of a tactile sensor for underwater manipulation. Water turbidity in energetic underwater environments can degrade the performance of perception sensors, making the execution of already difficult manipulation tasks even more challenging. Tactile sensing can provide useful information in these environments. One popular type of tactile sensor for terrestrial applications uses barometric pressure sensors encased in a soft elastomer. However, the performance of these sensors in changing ambient pressures has not been investigated. We designed a custom testbed to characterize high-pressure MEMS barometers embedded in two types of silicone up to 50 PSIG ambient pressure. Using characterization results from a single barometer, we then designed two 2,x,4 tactile grids. Datasets of differential pressures (against a control sensor) for varying contact locations were used to train feedforward neural networks for point load estimation. Results show that for the grid encased in softer silicone, the model performance improved as the ambient pressure increased (average RMSE of 0.33 mm).

Keywords: Soft Robot Applications, Soft Sensors and Actuators, Modeling, Control, and Learning for Soft Robots
Abstract: Elephants and other animals heavily rely on the sense of smell to operate. Soft robots would also benefit from an artificial sense of smell, which could be helpful in typical soft robotic tasks such as search and rescue, pipe inspection, and all the tasks involving unstructured environments. This work proposes an artificial nose on a soft robotic arm that ensures separate smell concentration readings. We propose designing the nose to generate a one-to-one matching between the sensors’ inputs and the actuators. This design choice allows us to implement a simple control strategy tailored to reach a dynamically varying smell in the environment, which we validate on a two-segment tendon-driven soft robotic arm equipped with the proposed artificial nose. We also propose and validate in simulation a control strategy for reaching tasks in the case of a stationary smell.

Keywords: Soft Sensors and Actuators, Medical Robots and Systems, Modeling, Control, and Learning for Soft Robots
Abstract: Enabling dexterous manipulation and safe human-robot interaction, soft robots are widely used in numerous surgical applications. One of the complications associated with using soft robots in surgical applications is reconstructing their shape and the external force exerted on them. Several sensor-based and model-based approaches have been proposed to address the issue. In this paper, a shape sensing technique based on Electrical Impedance Tomography (EIT) is proposed. The performance of this sensing technique in predicting the tip position and contact force of a soft bending actuator is highlighted by conducting a series of empirical tests. The predictions were performed based on a data-driven approach using a Long Short-Term Memory (LSTM) recurrent neural network. The tip position predictions indicate the importance of using EIT data along with pressure inputs. Changing the number of EIT channels, we evaluated the effect of the number of EIT inputs on the accuracy of the predictions. The least RMSE values for the tip position are 3.6 and 4.6 mm in Y and Z coordinates, respectively, which are 7.36% and 6.07% of the actuator's total range of motion. Contact force predictions were conducted in three different bending angles and by varying the number of EIT channels. The results of the predictions illustrated that increasing the number of channels contributes to higher accuracy of the force estimation. The mean errors of using 8 channels are 7.69%, 2.13%, and 2.96% of the total force range in three different bending angles.

Keywords: Soft Robot Applications, Soft Robot Materials and Design, Soft Sensors and Actuators
Abstract: Soft actuators exhibiting versatile behaviors have potential applications in robotics. This paper proposes kinematics, kinetics, and dynamic models of an online-programmable soft actuator. The actuator is composed of four strings and an inflatable textile tube folded inside a housing structure. Each string is controlled by a single DC motor which has an optical encoder. Pulling a string produces bending in one direction, while pulling the four strings in a coordinated manner produces additional motions. With the proposed forward and inverse kinematic model, the actuator was able to follow a desired end-effector trajectory in the Cartesian space. Furthermore, due to the dynamic model, our simulation study shows that the soft actuator can handle external force changes at the end-effector, such as mass changes and friction forces.

Keywords: Soft Robot Materials and Design, Soft Sensors and Actuators, Optimization and Optimal Control
Abstract: Design optimization of soft actuators is essential for task-oriented applications. Models derived from analytical solutions, the Finite Element Method (FEM), or empirical characterized datasets are widely used to estimate the response of the actuators during actuation, acting as the backbone for design optimization. Faced with the trade-off between speed and accuracy, substantial challenges occur when moving from simulation to optimization due to the compliant, high degree of freedom, and high-dimensional design space of the soft-bodied robot. FEM becomes increasingly computationally expensive with increased design complexity during optimization iterations, while the data-driven modeling approach (e.g., Artificial Neural Network) consumes significant resources prior to optimization. To address the challenge of highly nonlinear and non-convex design optimization in soft robots using the black box modeling, this paper compares of Bayesian optimization (BO) algorithm and genetic algorithm (GA) with FEM and Artificial Neural Network (ANN) models. The shape-matching of a multi-legged robot (a starfish) is demonstrated as an example of a task-oriented design scenario that presents design optimization challenges of the design space scalability. The experimental results show that the bi-level BO outperforms BO with FEM by achieving 2.8 to 9.8 times smaller objective values within a certain time for low-dimensional design problems; GA with the ANN model can achieve lower objective values 3 to 18 times faster in high-dimensional design problems than bi-level BO with FEM in low-dimensional design problems.

Keywords: Haptics and Haptic Interfaces, Soft Sensors and Actuators, Biologically-Inspired Robots
Abstract: The softness perception of objects with lower stiffness than that of robotic skin is challenging, as the proportion of the deformation of skin to that of an object's surface is unknown. This makes it difficult to derive the indentation depth typically used for stiffness estimation. To overcome this challenge, we implemented human-inspired softness sensing in a soft anthropomorphic finger based on tactile information alone without using the information about indentation depth or displacement. In the experiments where LSTM networks were trained to discriminate viscoelastic soft objects, we demonstrated that the sensorized robotic finger using tactile information from barometric sensors embedded in its soft skin could successfully learn to discriminate soft objects. By dissociating the relative contribution of the dynamic pattern of pressure distribution and that of local pressure, we further investigated how differences in available tactile information could impact the ability to distinguish the softness of viscoelastic objects. The results demonstrated that the pressure distribution and its change on the soft contact area of the robotic finger provided information to discriminate the softness of viscoelastic objects and that the tactile information about softness was spatiotemporal in nature. The results further implied that nonlinear local dynamics such as hysteresis in local pressure changes can provide additional information about the viscoelasticity of touched objects.

Keywords: Soft Robot Materials and Design, Modeling, Control, and Learning for Soft Robots, Soft Sensors and Actuators
Abstract: Untethered soft robots have great potential in applications ranging from search-and-rescue to human-assistive robotics, and the light weight, impact resistance, and innate mechanical intelligence of soft robotics would provide untethered soft robots with unique capabilities compared to traditional robotics. Despite their great potential, most soft robots are still tethered to their power sources and the few existing untethered platforms suffer from either slow motions (for pump-based systems) or short lifetimes and a lack of controllability (for propellant-based systems). In this work, we introduce the concept of a pump-controlled propellant-powered (PCPP) system in which a pump moves fuel into a reaction chamber, where the produced gasses can pressurize a soft pneumatic system. We present a model to compare the performance of a pneumatic and PCPP system, demonstrating the PCCP system’s favorable work savings and actuation speed. We then perform preliminary tests on a prototype system to validate the model, also demonstrating that the platform can inflate a soft actuator. In the future, the PCPP system has the potential to combine the best features of existing pneumatic and propellant systems, allowing for both controlled and fast-moving untethered soft robotic motion.

Keywords: Modeling, Control, and Learning for Soft Robots, Force and Tactile Sensing
Abstract: The compliant nature of vine robots, or everted tubes, makes them very useful for passively navigating sensitive and cluttered environments. Like many soft robots, their compliance makes kinematic modeling difficult. This work seeks to validate the assumption that everted tubes can be modeled as an inflated cantilever beam. The tip deflection and curvature of tubes constructed from Low Density Polyethylene (LDPE) and commercially available silicone coated nylon, two commonly used materials for everted tube robots, were measured under several different loading conditions in both everted and non-everted configurations. Our data show that everted tubes constructed from LDPE can be modeled with Comer's model for inflated isotropic beams, which has been assumed in recent literature. In contrast, this model did not fit the silicone coated nylon beams' behavior. This is because of the anisotropic properties of fabric the current model does not consider. Results also showed that everted tubes collapsed before the modeled maximum loading condition at a constant tip displacement, whereas uneverted tubes collapsed slightly after the modeled maximum load. From the inflated beam model, we present an iterative approach to computing the forward kinematics of an everted tube in contact with an obstacle within an environment. This allows for estimating contact forces and determining the static pose of an everted tube.

Keywords: Compliant Joint/Mechanism, Force Control, Grippers and Other End-Effectors
Abstract: In soft robotics, variable stiffening is the key to taking full advantage of properties such as compliance, manipulability and deformability. However, many variable stiffness actuators and mechanisms which have been produced so far to control these properties of soft robots are slow, bulky, or require additional complex actuators. This paper presents a novel passive soft joint based upon the intrinsic non-Newtonian behavior of Shear Thickening Fluids (STFs). The joint stiffness is varied by changing the speed at which it is actuated. The joints fabricated for testing have a simple cylindrical structure comprised of a soft silicone shell filled with a STF. Three prototypes with lengths of 20, 40 and 60mm were produced for experimental validation. We characterize the behavior of the joints in compression, expansion and bending, yielding a stiffness variation of more than 5x based on actuation speed in compression testing. This paper is the first step in producing a new category of variable stiffening mechanisms based on STFs which can be incorporated into soft robots without the need for additional actuation. It is envisaged that this new soft joint will find applications in soft manipulators and wearable devices.

Keywords: Soft Robot Materials and Design, Soft Robot Applications, Simulation and Animation
Abstract: Soft robotic grasping has rapidly spread through the academic robotics community in recent years and pushed into industrial applications. At the same time, multimaterial 3D printing has become widely available, enabling the monolithic manufacture of devices containing rigid and elastic sections. We propose a novel design technique that leverages both technologies and can automatically design bespoke soft robotic grippers for fruit-picking and similar applications. We demonstrate the novel topology optimisation formulation that generates multi-material soft grippers, can solve internal and external pressure boundaries, and investigate methods to produce air-tight designs. Compared to existing methods, it vastly expands the searchable design space while increasing simulation accuracy.

Keywords: Soft Robot Materials and Design, Additive Manufacturing, Soft Sensors and Actuators
Abstract: Soft underwater robots typically explore bioinspired designs at the expense of power efficiency when compared to traditional underwater robots, which limits their practical use in real-world applications. We leverage a fluidic closed-loop controller to actuate a passive underwater glider. A soft hydrostatic pressure sensor is configured as a bangbang controller actuating a swim bladder made from silicone balloons. Our underwater glider oscillates between the water surface and 4 m depth while traveling 15 m translationally. The fluidic underwater glider demonstrates a power efficiency of 28 mW/m. This work demonstrates a low-cost and powerefficient underwater glider and non-electronic controller. Due to its simple design, low cost, and ease of fabrication using FDM printing and soft lithography, it serves as a starting point for the exploration of non-electronic underwater soft robots.

Keywords: Grippers and Other End-Effectors, Grasping, Soft Robot Materials and Design
Abstract: A widespread methodology to enhance the design of robotic devices is represented by topology optimization. Typically, the optimization aims at designing a certain part of the robot to satisfy a priori, user-defined mechanical properties while minimizing the used material for building the structure. In this paper, we apply topology optimization to robotic grippers, and we propose to define the requirements for the optimization in a data-driven way based on simulated experiments of grasping tasks. Specifically, the architecture we propose is composed of three sequential phases. The input of the architecture includes the initial model of the gripper, the specific gripper component to be optimized, and a set of parameters. The first part of the architecture acquires force signals from the gripper component that are sensed during the grasping simulations. Hence, these signals are fed into the second phase, which analyzes the forces through pixel connectivity and Dynamic Time Warping algorithms and provides the instructions for the topology optimization. Ultimately, the third block performs the optimization. The method is tested by optimizing a specific part of a soft-rigid gripper. Results from simulation confirm that the proposed architecture provides an improved version of the original gripper, not only in terms of optimized use of materials but also in terms of grasp success rate.

Keywords: Soft Sensors and Actuators, Sensor-based Control, Modeling, Control, and Learning for Soft Robots
Abstract: 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.

Keywords: Soft Robot Materials and Design, Grippers and Other End-Effectors, Modeling, Control, and Learning for Soft Robots
Abstract: Multistable structures are characterized by the existence of more than one statically stable state, which can provide a reference point for open-loop control schemes leveraging these systems' intrinsic mechanics. Multistable soft robots can thus take advantage of both the adaptability of soft robotics and the mechanical response of multistable elements for the potential simplification of robotic control and predictability. We present an energy-based analytical model for a class of soft multistable grippers enabling the design and prediction of their stable states abstracted as programmed operational points. The analytical model based on lumped parameter springs allows us to predict the system's final state upon actuation with reduced computational time compared to Finite Element (FE) simulations. The obtained computational efficiency enables us to search the configuration space in a tractable fashion, thereby facilitating the rational design of our grippers' set points. We validate our model against FE simulations and experimental tests. The model captures the fundamental mechanics of the introduced soft gripper topology, laying the foundation for efficient design optimization and simplified control of soft robots.

Keywords: Force Control, Compliant Joint/Mechanism, Manipulation Planning
Abstract: When humans perform dynamic motions such as throwing, the passive properties such as the stiffness and damping of their arm is known to contribute to the task performance. By developing a robot arm which enables the stiffness of the different joints to be set programmatically, its contribution to the throwing behaviours can be determined. In addition to enabling new capabilities in robots this can also be useful for understanding how humans may perform such tasks. Utilizing permanent magnet synchronous motors (PMSM) and integrating them in back-drivable configurations we present a method of achieving programmable, precise, high bandwidth stiffness control. With a two joint variable stiffness arm, we experimentally explore the role of stiffness and coordination of actuation timings for the throwing of a Frisbee disk. From this exploration key trends between stiffness and the throwing distance and angle are observed. Considering variable stiffness (VS) we also see that the role and significance of VS varies depending on the overall energy levels of the system. For low energies, having a constant torque profile can enable a 30% increase in throwing distance, where as at higher energies VS is less significant. When compared to human throwers, the robot performs comparable to experienced humans for a short distance throwing task.

Keywords: Telerobotics and Teleoperation, Legged Robots, Modeling, Control, and Learning for Soft Robots
Abstract: Soft robotics holds tremendous potential for various applications, especially in unstructured environments such as search and rescue operations. However, the lack of autonomy and teleoperability, limited capabilities, absence of gait diversity and real-time control, and onboard sensors to sense the surroundings are some of the common issues with soft-limbed robots. To overcome these limitations, we propose a spatially symmetric, topologically-stable, soft-limbed tetrahedral robot that can perform multiple locomotion gaits. We introduce a kinematic model, derive locomotion trajectories for different gaits, and design a teleoperation mechanism to enable real-time human-robot collaboration. We use the kinematic model to map teleoperation inputs and ensure smooth transitions between gaits. Additionally, we leverage the passive compliance and natural stability of the robot for toppling and obstacle navigation. Through experimental tests, we demonstrate the robot's ability to tackle various locomotion challenges, adapt to different situations, and navigate obstructed environments via teleoperation.

Keywords: Soft Robot Applications, Modeling, Control, and Learning for Soft Robots, Biologically-Inspired Robots
Abstract: In this work, we present a whole-arm grasping strategy for soft arms whose task is to capture space debris. The non-cooperative nature of space debris and the characteristics of the space environment enforce high-level requirements for robotic arms, especially dexterity. Taking inspiration from the outstanding capabilities of the elephant trunk in grasping, we formulated a grasping strategy based upon the identification of contact points on the object to force the bending of the arm and induce the wrapping around the object, as the animal model does. This strategy is implemented by leveraging on coupled Finite Element simulations of a trunk-like soft arm and Reinforcement Learning tools to learn the grasping. The results show that the robot successfully learns the task by moving the proximal part closer to the object and devoting the distal one to the wrapping around the object. We show that the obtained policy is valid for diverse object sizes and positions. Our grasping strategy is the first example of bio-inspired whole-arm grasping for a soft arm in space. We believe that, in the near future, this strategy will enable new grasping capabilities in soft arms.

Keywords: Cellular and Modular Robots, Biologically-Inspired Robots, Modeling, Control, and Learning for Soft Robots
Abstract: In nature, cells combine into different structures to perform required tasks and can break and rejoin to make smaller and larger organisms. Taking inspiration from cells, we present an adaptive soft robot composed of simple modular elements (cells) in a linear arrangement, joined together by magnets, capable of performing locomotion by exploiting frictional asymmetries with the terrain. Using a simple control mechanism to change their volumetric actuation, a travelling wave was generated to move the robot. Based on the inflation profile of the cell, we defined 4 geometric states, S1 (contracted), S2 (relaxed state), S3 (intermediate state) and S4 (inflated). In locomotion gaits, each cell can act as a foot or a muscle, depending on degree of inflation, and change function throughout the gait. The modular robot can also separate itself into multiple parts and recombine as needed, demonstrating attractive capabilities for autonomous exploration in natural environments. This can potentially be used to remove damaged cells or change the shape of the robot body. We present the design of the modular soft robot and demonstrate its locomotion and reconfiguration capabilities.

Keywords: Modeling, Control, and Learning for Soft Robots, Surgical Robotics: Steerable Catheters/Needles, Force and Tactile Sensing
Abstract: Recent advances on Concentric Tube Robots (CTRs) enable the construction and analysis of concentric combinations of precurved elastic tubes. These robots are very appropriate for performing Minimally Invasive Surgery (MIS) with a reduction in patient recovery time. In this work, we propose a kinetostatic model for CTRs based on the Geometric Variable-Strain (GVS) approach where the tubes’ sliding motion, the distributed external forces along the tubes and concentrated external forces at the tip, are included. Our approach allows us to estimate the shape of CTRs and the tip forces using the displacements of the tubes and the insertion and rotation input forces and torques. Moreover, we propose a modification in the model, which eliminates completely the sliding friction among the tubes. This new approach opens a new way to use CTRs in surgical applications without the need of sensors along the tubes, but only actuation measurements. The simulation results demonstrate the effectiveness of the proposed approach.

Keywords: Additive Manufacturing, Modeling, Control, and Learning for Soft Robots, Tendon/Wire Mechanism
Abstract: Construction 3D printing technology has recently received significant attention as a method for creating construction components or printing entire buildings. The deployment of Cable Driven Parallel Robots (CDPRs) in large-scale 3D printing is being explored as a potential candidate due to their low cost, high speed, and design modularity. However, the cable's inertial and elastic properties may lead to sagging and vibration, making the system difficult to model. In this paper, we use the Geometric Variable Strain (GVS) model, a geometrically exact approach based on the Cosserat rod theory, to model the dynamics of a CDPR. The Cosserat rod theory accounts for deformation modes that are not considered in other models, while the geometric formulation ensures accurate and fast computation. We compare the dynamic simulation of a small-scale CDPR prototype at different speeds and with an experimental setup. We also study the dynamics of a large-scale system subject to step loading. We show that analyses of CDPR systems using the GVS approach can reveal new perspectives on their control, design, and development.

Keywords: Modeling, Control, and Learning for Soft Robots, Grasping, Soft Robot Applications
Abstract: In this paper, an integrated design of a soft gripper is described for an efficient mushrooms harvesting. The soft gripper is made up multi-phalanges soft fingers in order to address the shape adaptability issues regarding the form enclosure grasping strategy. The shape kinematics of these soft fingers has been described using parametric curves, namely the Pythagorean Hodograph (PH) curves, with a prescribed length. This has enabled a Reduced Order Modeling (ROM) by using a few number of geometric control points. Then, Euler-Bernoulli (EB) modeling technique has been applied to these curves to estimate the actuation control inputs, allowing the mushrooms to be grasped under optimal safety conditions. The real-time grasping control issues based on the sliding Mode, have been discussed using a combined action of the attractive and repulsive Artificial Potential Field (APF), used to drive the soft gripper to the mushroom target. This control has been applied to the virtual control points of their representative PH curves, and yielded an accurate positioning of the soft gripper during the grasping process. The safety and the quality of the mushroom during the harvesting has been guaranteed by the presence of the contact force sensors, as well as the hyper-elastic material constituting each soft finger. The above strategy keeps the harvested mushroom safe during the grasping and therefore, enables a real-time shape control for a form enclosure soft grasping. The results of the proposed technique have been experimentally assessed using a 3-fingers soft gripper made up of Fluidic Elastomeric Actuators (FEAs) in an agriculture fresh mushrooms farm.

Keywords: Soft Robot Materials and Design, Modeling, Control, and Learning for Soft Robots, Deep Learning in Robotics and Automation
Abstract: Soft robot designs based on auxetic lattices could offer superior dexterity, tunable local kinematics, and morphological intelligence. However, the design and control of these structures for robotic tasks, requiring multiple states and motions, remains a challenging problem. Finite element models (FEMs) offer a promising way of predicting robot behaviour that might be used for design and control optimization. Yet, these physics-based models often have high computational cost and can not provide explicit gradient information to guide the search for optimal designs. In this paper, we abstract the physical predictions of FEMs through differentiable surrogate models and demonstrate design and trajectory optimization using a gradient-based optimizer. We compare the performance of convolutional neural networks (CNNs) and graph neural networks (GNNs) as surrogate models. We then demonstrate the use of a gradient-based optimizer to find optimal designs for a specified deformation and optimal pairs of designs and actuation inputs for a trajectory specified by waypoints. In each case, the differentiable surrogate model enables the gradient-based optimizer to discover novel designs lying outside of the training data that achieve the required motions (with an error less than 5.83mm for deformations and less than 1.26mm for trajectories).

Keywords: Soft Sensors and Actuators, Soft Robot Materials and Design, Soft Robot Applications
Abstract: Detection and control of the physical contact/impact between micro aerial vehicles and the surrounding obstacles have become a significant issue with the rapid growth of their use in inspection and mapping missions in confined, obstacle-cluttered environments. In this work, we introduce a collision-resilient compliant micro quadcopter equipped with soft coil-spring type force sensors to passively resist and detect the physical contact/impact of the drone. The sensors act as resistive elements with a nominal resistance of 130-150 kOhm. They are manufactured from a conductive material via FDM 3D printing. We install these sensors on the protective bumpers of the collision-resilient foldable body of the drone. Any contact/impact between the bumpers and an obstacle results in deformation and buckling of the soft sensors, which results in a drastic change in their resistance, making it possible to detect the contacts/impacts of the bumpers. With a total weight of 220g and dimensions of 22cm×22cm×9cm, SCoReR successfully detects and recovers 100% of the contacts/impacts when it approaches a rigid wall with a velocity in the range of [0.1-1] m/s.

Keywords: Grippers and Other End-Effectors, Modeling, Control, and Learning for Soft Robots, Human Factors and Human-in-the-Loop
Abstract: Robot arms that assist humans should be able to pick up, move, and release everyday objects. Today's assistive robot arms use rigid grippers to pinch items between fingers; while these rigid grippers are well suited for large and heavy objects, they often struggle to grasp small, numerous, or delicate items (such as foods). Soft grippers cover the opposite end of the spectrum; these grippers use adhesives or change shape to wrap around small and irregular items, but cannot exert the large forces needed to manipulate heavy objects. In this paper we introduce RIgid-SOft (RISO) grippers that combine switchable soft adhesives with standard rigid mechanisms to enable a diverse range of robotic grasping. We develop RISO grippers by leveraging a novel class of soft materials that change adhesion force in real-time through pneumatically controlled shape and rigidity tuning. By mounting these soft adhesives on the bottom of rigid fingers, we create a gripper that can interact with objects using either purely rigid grasps (pinching the object) or purely soft grasps (adhering to the object). This increased capability requires additional decision making, and we therefore formulate a shared control approach that partially automates the motion of the robot arm. In practice, this controller aligns the RISO gripper while inferring which object the human wants to grasp and how the human wants to grasp that item. Our user study demonstrates that RISO grippers can pick up, move, and release household items from existing datasets, and that the system performs grasps more successfully and efficiently when sharing control between the human and robot. See videos here: https://youtu.be/5uLUkBYcnwg

Keywords: Soft Robot Applications, Soft Sensors and Actuators, Medical Robots and Systems
Abstract: Flexible ultrasound (US) transducer, which has a potential to fit various regions of human body for diagnosis, need to have its geometry accurately measured for US image reconstruction. This paper serves a shape sensing system with the electrostatic differential capacitance for the imaging with the flexible US transducer. The shape sensing system is composed of two strips as a pair, each end of which is fixed, and focuses on the relative shift between capacitance sensors embedded in the inner and outer strips when bending the sensing system. For increasing the capacitance, we applied a silicon oil to the sensor substrate and changed the size of electrodes. Experimental results showed that the estimation error was improved by the average of 52.8% when applying the silicon oil and the average of 10.4% by increasing the size of electrodes. Additionally, US simulation was performed for investigating the influence of image reconstruction due to the sensing error. The simulation results enabled to visualize all point targets and demonstrated the feasibility that the developed sensing system are applicable for the flexible US transducer.

Keywords: Soft Robot Applications
Abstract: Acoustic signals can be used to detect environmental interactions of everting tube robots. This experiment distinguishes differences in pressure and audio signals in tubes freely everting through different-sized tunnels, with acoustic signal measurement ranging from 0-10 kHz. Pressure rises when transitioning to smaller tunnels and drops when transitioning to larger tunnels. Audio becomes louder when transitioning to larger tunnels and quieter when transitioning to smaller tunnels. Audio FFTs and spectrograms also show distinguishable eversion sounds and clear evidence of tunnel transitions. Time data suggests that reliable time series models could be created to detect tunnel transitions. Frequency data also suggests that a reliable image-analysis model could be created to detect tunnel transitions.

Keywords: Soft Sensors and Actuators, Perception for Grasping and Manipulation, Modeling, Control, and Learning for Soft Robots
Abstract: Soft robotics is applicable to a variety of domains due to the adaptability offered by the soft and compliant materials. To develop future intelligent soft robots, soft sensors that can capture deformations with nearly infinite degrees of freedom are necessary. Soft sensor networks can address this problem, however, measuring all sensor values throughout the body requires excessive wiring and complex fabrication that may hinder robot performance. We circumvent these challenges by developing a non-invasive measurement technique, which is based on an algorithm that solves the inverse problem of resistor network, and implement this algorithm on a soft resistive, strain sensor network. Our algorithm works by iteratively computing the resistor values based on the applied boundary voltage and current responses, and we analyze the reconstruction error of the algorithm as a function of network size and measurement error. We further develop electronics setup to implement our algorithm on a stretchable resistive strain sensor network made of soft conductive silicone, and show the response of the measured network to different deformation modes. Our work opens a new path to addressing the challenge of measuring many sensor values in soft sensors, and could be applied to soft robotic sensor systems.

Keywords: Multifingered Hands, Force and Tactile Sensing, Soft Sensors and Actuators
Abstract: The use of tactile sensing exhibits benefits over visual detection as it can be deployed in occluded environments and can provide deeper information about an object’s material properties. Soft hands have increasingly been used for tactile object identification, providing a high degree of adaptability without requiring complex control schemes. In this work, we propose a framework for identifying a range of objects in any pose by exploiting the compliance of a soft hand equipped with distributed tactile sensing. We propose EDAMS, an Encoder-Decoder Architecture for Multi-grasp Soft sensing and an ad-hoc data structure capable of encoding information on multiple grasps, while decoupling the dependency on the pose order. We train the model to map the high-dimensional multi-grasp tactile sensor data into a lower-dimensional latent space capable of achieving the geometrical separation of each object class, and enabling accurate object classification. We provide an empirical analysis of the benefit of multi-grasp perception for object identification, and show its impact on the separation of the objects in sensor space. Notably, we find the classification accuracy to change widely across the number of grasps, ranging from 47.0% for a single grasp, to 99.9% for 10 grasps.

Keywords: Modeling, Control, and Learning for Soft Robots, Cellular and Modular Robots, Legged Robots
Abstract: Path-tracking is often challenging in miniature robots because their feet or wheels tend to slip due to the low robot weight. In this work, we investigate the effect of c-leg compliance on path-tracking performance and the obstacle-climbing capabilities of our foldable and miniature robot with soft, c-shaped legs. With its 82 mm x 60 mm x 29 mm size and 29.25 grams weight, a single module of our robot is one of the smallest untethered miniature robots. Our results show that utilizing soft c-shaped legs provides smooth path-tracking performance, similar to a wheeled differential drive robot. However, modules with rigid c-shaped legs are affected significantly by the impact and slip between the leg and the ground, and they perform rather unpredictably. Additionally, modules with wheels cannot climb obstacles 1 mm or larger. We show that using soft legs enhances the obstacle climbing skills of modules by climbing a 9 mm obstacle, while the module with rigid legs can only climb a 7 mm obstacle. These path-tracking abilities and obstacle-climbing capacity support our vision to build a reconfigurable robot using these modules.

Keywords: Rehabilitation Robotics, Sensor-based Control, Wearable Robots
Abstract: Stroke can result in limb disorders, and many stroke victims will experience unilateral limb problems. Therefore, post-disease rehabilitation, particularly for the upper limbs, is crucial. In general, professionals or family members must accompany patients during rehabilitation. However, it is inconvenient and could impede the development of the recovery process. Therefore, a rehabilitation tool that is self-controllable by the patient, basic in design, and simple to use is beneficial to the patient's recovery. In this study, to help stroke patients with unilateral limb abnormalities regain self-control, a set of elbow rehabilitation equipment was created. It is simple to use, safe, and reliable. The patient can control the soft robot to mend the elbow using the surface Electromyography (sEMG) signal from the healthy arm. Five classification models' classification results were compared, and Support Vector Machine (SVM) accuracy was found to be 97% accurate. The SVM model is chosen for real-time control, considering classification accuracy, stability, and operation time. The elbow flexion and extension real-time control experiments are conducted. The system execution time can be as long as 2.275 seconds and 2.188 seconds when elbow flexion and extension are performed, respectively. For the rehabilitation of stroke patients, this study provides a wide range of possible applications.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Materials and Design, Soft Robot Applications
Abstract: Soft, growing inflated beam robots, also known as everting vine robots, have previously been shown to navigate confined spaces with ease. Less is known about their ability to navigate three-dimensional open spaces where they have the potential to collapse under their own weight as they attempt to move through a space. Previous work has studied collapse of inflated beams and vine robots due to purely transverse or purely axial external loads. Here, we extend previous models to predict the length at which straight vine robots will collapse under their own weight at arbitrary launch angle relative to gravity, inflated diameter, and internal pressure. Our model successfully predicts the general trends of collapse behavior of straight vine robots. We find that collapse length increases non-linearly with the robot's launch angle magnitude, linearly with the robot's diameter, and with the square root of the robot's internal pressure. We also demonstrate the use of our model to determine the robot parameters required to grow a vine robot across a gap in the floor. This work forms the foundation of an approach for modeling the collapse of vine robots and inflated beams in arbitrary shapes.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Sensors and Actuators, Soft Robot Materials and Design
Abstract: Soft grippers have gained a lot of interest in the last decade. In addition to firmly grasping an object, the estimation of its hardness is also an important aspect in various soft robotic applications. This study proposes a shape-invariant indirect hardness estimation approach for a soft vacuum-actuated gripper with an embedded depth camera. The technique proposed herein would eliminate the need for invasive sensors, which may damage certain objects. The project focuses on a simultaneous grasping and sensing system for deformable objects, without visible markers on the gripper’s membrane. The deformation of membrane, containing valuable information on the object’s properties, is captured by a depth camera inside the gripper. A convolutional neural network-based hardness prediction model is created with a mean absolute percentage error (MAPE) of 0.37%, in the case of trained shapes and trained hardnesses. For untrained hardnesses, the error is observed to be 4.54%. Through comparison with conventional grayscale images, the experiments also showed that images with depth information are more preferable for hardness estimation.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Applications
Abstract: Soft robot are celebrated for their propensity to enable compliant and complex robot-environment interactions. Soft robotic manipulators, or slender continuum structure robots have the potential to exploit these interactions to enable new exploration and manipulation capabilities and safe human-robot interactions. However, the interactions, or perturbations by external forces cause the soft structure to deform in an infinite degree of freedom (DOF) space. To control such system, reduced order models are needed; typically models consider piecewise sections of constant curvature although external forces often deform the structure out of the constant curvature hypothesis. In this work we perform an analysis of the trade-off between computational treatability and modelling accuracy. We then propose a new kinematic model, the Piece-wise Affine Curvature (PAC) which we validate theoretically and experimentally showing that this higher-order model better captures the configuration of a soft continuum body robot when perturbed by the external forces. In comparison to the current state of the art Piece-wise Constant Curvature (PCC) model we demonstrate up to 30% reduction in error for the end position of a soft continuum body robot.

Keywords: Modeling, Control, and Learning for Soft Robots, Medical Robots and Systems, Soft Robot Applications
Abstract: The introduction of soft robots has led to the development of inherently safe and flexible interventional tools for medical applications, when compared to their traditionally rigid counterparts. In particular, robot-assisted surgery is one of the medical applications that benefits from the inherent properties of soft instruments. However, robust control and reliable manipulation of soft tools remain challenging. In this paper, we present a new method based on reduced finite elementmethod model and closed-loop inverse kinematics control for a fiber-reinforced soft robot. The highly flexible, pneumatically driven soft robot has three fully fiber-reinforced chamber pairs. The outer diameter is 11.5 mm. An inner working channel of 4.5 mm provides a free lumen for in-vivo cancer imaging tools during minimally invasive interventions. Here, a device is introduced through the inner free channel of the manipulator to retrieve a tissue biopsy which can then be investigated for cancerous tissue using SpiderMass technology. Simulation and experimental results are compared to validate the model and control methods, using one-module and two-module robots. The results show a real-time control is achievable using the reduced model. Combing the closed-loop control, the median position tracking errors are generally less than 2 mm.

Keywords: Modeling, Control, and Learning for Soft Robots, Learning and Adaptive Systems, Soft Robot Applications
Abstract: Recently, learning-based controllers that leverage mechanical models of soft robots have shown promising results. This paper presents a closed-loop controller for dynamic trajectory tracking with a pneumatic soft robotic arm learned via Deep Reinforcement Learning using Proximal Policy Optimization. The control policy was trained in simulation leveraging a dynamic Cosserat rod model of the soft robot. The generalization capabilities of learned controllers are vital for successful deployment in the real world, especially when the encountered scenarios differ from the training environment. We assessed the generalization capabilities of the controller in silico for four tests. The first test involved the dynamic tracking of trajectories that differ significantly in shape and velocity profiles from the training data. Second, we evaluated the robustness of the controller to perpetual external end-point forces for dynamic tracking. For tracking tasks, it was also assessed the generalization to similar materials. Finally, we transferred the control policy without retraining to intercept a moving object with the end-effector. The learned control policy has shown good generalization capabilities in all four tests.

Keywords: Soft Sensors and Actuators, Modeling, Control, and Learning for Soft Robots, Grasping
Abstract: Grasping control is one of the key features of robot manipulation. Slipping detection, avoidance, and minimum force grasping are of primary concern since it is expected that robot manipulators have similar performance to human hands. In this work, a new type of optoelectronic sensor, which has a human-like soft skin but a simple design, is applied to slip motion control. Based on the model of this soft sensor and the robotic gripper, we describe a model reference adaptive controller (MRAC) to estimate unknown system parameters for grasping random objects. Update laws for unknown parameters are chosen by stability analysis and the system feasibility is illustrated through both numerical simulation and hardware experiment.

Keywords: Contact Modeling, Simulation and Animation, Soft Sensors and Actuators
Abstract: Soft robots offer an exciting and novel alternative to traditional robots composed of rigid bodies. Many of the primary benefits soft robots have over more traditional robots result from their inherent compliance and their potential for low force interactions with their environments. Therefore, modeling soft robots requires the ability to accurately simulate contact mechanics. In this work, we present the solution of contact mechanics finite element problems specifically for soft robots in a MOOSE-based multiphysics simulation platform we developed, Kraken. The primary contributions of this work are threefold. Firstly, our implementations enable the modeling of additional types of contact critical to the simulation of soft robots. Next, we demonstrate how our new self contact method can be used to dramatically decrease the computational cost of contact modeling. Finally, we demonstrate the abilities of Kraken as a platform to simulate the complex interactions of soft robots and the environment.

Keywords: Deep Learning in Robotics and Automation, Learning and Adaptive Systems, Modeling, Control, and Learning for Soft Robots
Abstract: This paper presents an error-driven adaptive scheme for soft arm control. It consists of two components: an embedded control policy and an online regressing network (ORN). The embedded control policy is trained to achieve a desired task in a training environment, and the ORN learns to adjust the input to the control policy to enable its implementation with a physical soft robot. The ORN accomplishes this by utilizing the mismatch information between the training environment and the physical soft robot. The control policy learns to follow a desired trajectory with appropriate accuracy in offline mode with a data-driven dynamics model of the soft robot using soft actor critic (SAC) algorithm. The trained policy upon testing with the actual soft robot exhibits significant training-to-reality gap. The proposed control scheme learns to overcome this training-to-reality gap in a matter of seconds. Its adaptability is further tested by employing a previously trained policy with the modified performance of the robot under constant external stress. It is observed that the control scheme manages to adapt to the new robot setting without needing to retrain the policy from scratch to achieve desired accuracy. This method could be particularly advantageous for developing a control solution that can be broadly applied to account for the stochastic behavior of soft robots, which may arise due to factors such as material hysteresis, manufacturing inconsistencies, inaccuracies in external tracking systems, and variable initial conditions.

Keywords: Modeling, Control, and Learning for Soft Robots, Compliant Joint/Mechanism, Soft Sensors and Actuators
Abstract: Variable stiffness actuator (VSA) designs are manifold. Conventional model-based control of these nonlinear systems is associated with high effort and design-dependent assumptions. In contrast, machine learning offers a promising alternative as models are trained on real measured data and nonlinearities are inherently taken into account. Our work presents a universal, learning-based approach for position and stiffness control of soft actuators. After introducing a soft pneumatic VSA, the model is learned with input-output data. For this purpose, a test bench was set up which enables automated measurement of the variable joint stiffness. During control, Gaussian processes are used to predict pressures for achieving desired position and stiffness. The feedforward error is on average 11.5% of the total pressure range and is compensated by feedback control. Experiments with the soft actuator show that the learning-based approach allows continuous adjustment of position and stiffness without model knowledge.

Keywords: Simulation and Animation, Soft Robot Applications, Soft Robot Materials and Design
Abstract: Growing robots and their eversion principle have wide applications ranging from surgery to industrial inspection and archaeology. The eversion process involves deploying an inflatable device with a material located at the tip of the robot, which, when under pressure, elongates the robot's body. However, the simulation of this complex kinematic phenomenon is a significant challenge. Our approach proposes to use a combination of kinematics and quasi-static modeling to parameterize the starting conditions of the eversion process. This facilitates the understanding of the behavior of this complex kinematic phenomenon and help identify factors that have a significant impact on the eversion process and its response to external factors. The kinematic model uses the Cosserat rod models for local coordinates, while the quasi-static model is based on finite element analysis. The two models are combined to capture the behavior of the robot tip during eversion. This approach has been implemented and tested using the SOFA framework and has been evaluated on the deployment of a vine robot on a narrow passage. The results of our approach are encouraging to better understand the behaviour of soft growing robot during eversion.

Keywords: Soft Sensors and Actuators, Hydraulic/Pneumatic Actuators
Abstract: Soft pneumatic actuators are widely used for soft grippers, which are known for their compliance as compared with traditional grippers. The generated force/torque of soft pneumatic actuators directly determines the grasping force. This paper introduces a computationally efficient soft pneumatic actuator (SPA) design methodology. The complex structure of the pneumatic actuator is approximated by a cantilever beam. The relationship between input pressure and output torque is derived by standard mechanical analysis. The design problem is formulated as a model-based optimization problem by treating the input-output mathematical model as the objective function. By solving the optimization problem, the optimal design parameters are obtained. Finite element analysis is applied to preliminarily verify the design parameters without the time-consuming fabrication of many actuators. Three soft actuators with different design parameter sets were fabricated to validate the optimal parameters. This work shows the utility of surprisingly simple calculations and assumptions for rapid parametric design studies.

Keywords: Soft Robot Materials and Design, Soft Robot Applications, Hydraulic/Pneumatic Actuators
Abstract: Periodic inspection of aging gas pipes is important. However, the conventional inspection approach of excavation is unfriendly to the environment. From the perspective of sustainable development goals (SDGs), in this study, we introduced a pneumatically driven robot system called WATER-7 to observe the inner environment of aging pipes, in particular water inside these pipes, without excavation. The robot can locomote similar to an inchworm with a thrust module operating in a periodical pattern, select direction with an active bending module, and acquire images using a camera. The robot is designed and assembled within a diameter of 12[mm] to enable insertion into a gas meter valve as well as transition and retrieval from a 7[m] service pipe consisting of 8 pipe bends. To adapt to the general use environment, we miniaturized the robot, shortened the transit time by increasing air flow, and improved the robustness of each component. Furthermore, an algorithm for autonomous burr avoidance using an image analysis software was developed. According to experiments, the robot average transit time and retrieval without damage count for the assumed scenario were 81[min] and 9 times, respectively. In addition, the autonomous burr avoidance was confirmed to be effective.

Keywords: Soft Sensors and Actuators, Soft Robot Materials and Design, Force and Tactile Sensing
Abstract: Robots have been brought to work close to humans in many scenarios. For coexistence and collaboration, robots should be safe and pleasant for humans to interact with. To this end, the robots could be both physically soft with multimodal sensing/perception, so that the robots could have better awareness of the surrounding environment, as well as to respond properly to humans' action/intention. This paper introduces a novel soft robotic link, named emph{ProTac}, that possesses multiple sensing modes: tactile and proximity sensing, based on computer vision and a functional material. These modalities come from a layered structure of a soft transparent silicon skin, a polymer dispersed liquid crystal (PDLC) film, and reflective markers. Here, the PDLC film can switch actively between the opaque and the transparent state, from which the tactile sensing and proximity sensing can be obtained by using cameras solely built inside the ProTac link. In this paper, inference algorithms for tactile proximity perception are introduced. Evaluation results of two sensing modalities demonstrated that, with a simple activation strategy, ProTac link could effectively perceive useful information from both approaching and in-contact obstacles. The proposed sensing device is expected to bring in ultimate solutions for design of robots with softness, whole-body and multimodal sensing, and safety control strategies.

Keywords: Force and Tactile Sensing, Soft Sensors and Actuators, Grippers and Other End-Effectors
Abstract: Camera-based tactile sensors have shown great promise in enhancing a robot's ability to perform a variety of dexterous manipulation tasks. Advantages of their use can be attributed to the high resolution tactile data and 3D depth map reconstructions they can provide. Unfortunately, many of these tactile sensors use either a flat sensing surface, sense on only one side of the sensor's body, or have a bulky form-factor, making it difficult to integrate the sensors with a variety of robotic grippers. Of the camera-based sensors that do have all-around, curved sensing surfaces, many cannot provide 3D depth maps; those that do often require optical designs specified to a particular sensor geometry. In this work, we introduce GelSight360, a fingertip-like, omnidirectional, camera-based tactile sensor capable of producing depth maps of objects deforming the sensor's surface. In addition, we introduce a novel cross-LED lighting scheme that can be implemented in different all-around sensor geometries and sizes, allowing the sensor to easily be reconfigured and attached to different grippers of varying DOFs. With this work, we enable roboticists to quickly and easily customize high resolution tactile sensors to fit their robotic system's needs.

Keywords: Soft Sensors and Actuators, Wearable Robots, Medical Robots and Systems
Abstract: Smart soft wearable devices have great potential to change how technology is integrated into daily life. A particularly impactful and growing application is continuous medical monitoring; being able to stream physiological and behavioral information creates personalized datasets that can lead to more tailored treatments, diagnoses, and research. An area that can greatly benefit from these developments is lymphedema management, which aims to prevent a potentially irreversible swelling of limbs due to causes such as breast cancer surgeries. Compression sleeves are the state of the art for treatment, but many open questions remain regarding effective pressure and usage prescriptions. To help address these, this work presents a soft pressure sensor, a way to integrate it into wearable devices, and sensorized compression sleeves that continuously monitor pressure and usage. There are significant challenges to developing sensors for high-pressure applications on the human body, including operating between soft compliant interfaces, being safe and unobtrusive, and reducing calibration for new users. This work compares two sensing approaches for wearable applications: a custom pouch-based pneumatic sensor, and a commercially available resistive sensor. Experiments systematically explore design considerations including sensitivity to ambient temperature and pressure, characterize sensor response curves, and evaluate expected accuracies and required calibrations. Sensors are then integrated into compression sleeves and worn for over 115 hours spanning 10 days.

Keywords: Soft Robot Applications
Abstract: Tensegrity structures have been actively studied in recent years because they are lightweight, compliant, and flexible, which are properties not typically found in conventional robots. This structure can be modularized to create soft robots that operate in unknown environments such as cave or space with more complex and effective behavior. The basic deformation elements in modularization are stretching, bending, and torsion. Among them, torsional motion is important for proper manipulation and rotational operation. However, active, and large torsion in soft tensegrity structures has not been developed. Therefore, this study describes torsional deformation and a novel arrangement method for thin artificial muscles. The proposed method leads to the optimal placement of artificial muscles for torsion, by which we generated a large torsion of plus/minus 50 deg. This is more than 2.5 times larger than that of a previous tensegrity without compromising the favorable properties of the structure. Furthermore, by modularizing the tensegrity structure, a tensegrity arm capable of removing a plastic bottle cap was developed. The applicability of torsional deformation and the usefulness of modularization of the structure are demonstrated.

Keywords: Soft Sensors and Actuators, Soft Robot Materials and Design
Abstract: With infinite degrees of freedom, soft robots are expected to achieve dexterous and complex tasks, but this also puts forward higher requirements for their sensing capabilities. An important sensing task in soft robots is sensing their own deformation and current shape. Currently, most of the existing soft shaping sensors are limited by local perception abilities, stretchability, and fabrication difficulties. We propose a sensing method based on Electrical Impedance Tomography (EIT), which reconstructs conductivity patterns distributed on a surface, by considering the deformation-caused resistance changes. Comparison between the theoretical and experimental patterns reveals that even though the quality of the pattern is affected by a large amount of noise, the considered features are still able to reflect the change of shape. With the help of neural networks, the pattern is decoded to the physical data related to the deformation. Detection of the planar shape changes and proprioception of a sensor-integrated soft robot are presented to exhibit the capability of our method. Results show that the detected error ratios are mostly under 5% and 3% for 2D and 3D conditions respectively.