Learning-Based Control for Soft Robot-Environment Interaction with Force/Position Tracking Capability

Soft robotics promises to achieve safe and efficient interactions with the environment by exploiting its inherent compliance and designing control strategies. However, effective control for the soft robot-environment interaction has been a challenging task. The challenges arise from the nonlinearity and complexity of soft robot dynamics, especially in situations where the environment is unknown and uncertainties exist, making it difficult to establish analytical models. In this study, we propose a learning-based optimal control approach as an attempt to address these challenges, which is an optimized combination of a feedforward controller based on probabilistic model predictive control and a feedback controller based on nonparametric learning methods. The approach is purely data-driven, without prior knowledge of soft robot dynamics and environment structures, and can be easily updated online to adapt to unknown environments. A theoretical analysis of the approach is provided to ensure its stability and convergence. The proposed approach enabled a soft robotic manipulator to track target positions and forces when interacting with a manikin in different cases. Moreover, comparisons with other data-driven control methods show a better performance of our approach. Overall, this work provides a viable learning-based control approach for soft robot-environment interactions with force/position tracking capability.

Zero-shot model-free learning of periodic movements for a bio-inspired soft-robotic arm

In recent years, soft robots gain increasing attention as a result of their compliance when operating in unstructured environments, and their flexibility that ensures safety when interacting with humans. However, challenges lie on the difficulty to develop control algorithms due to various limitations induced by their soft structure. In this paper, we introduce a novel technique that aims to perform motion control of a modular bio-inspired soft-robotic arm, with the main focus lying on facilitating the qualitative reproduction of well-specified periodic trajectories. The introduced method combines the notion behind two previously developed methodologies both based on the Movement Primitive (MP) theory, by exploiting their capabilities while coping with their main drawbacks. Concretely, the requested actuation is initially computed using a Probabilistic MP (ProMP)-based method that considers the trajectory as a combination of simple movements previously learned and stored as a MP library. Subsequently, the key components of the resulting actuation are extracted and filtered in the frequency domain. These are eventually used as input to a Central Pattern Generator (CPG)-based model that takes over the generation of rhythmic patterns at the motor level. The proposed methodology is evaluated on a two-module soft arm. Results show that the first algorithmic component (ProMP) provides an immediate estimation of the requested actuation by avoiding time-consuming training, while the latter (CPG) further simplifies the execution by allowing its control through a low-dimensional parameterization. Altogether, these results open new avenues for the rapid acquisition of periodic movements in soft robots, and their compression into CPG parameters for long-term storage and execution.

Recent Synergies of Machine Learning and Neurorobotics: A Bibliometric and Visualized Analysis

Over the past decade, neurorobotics-integrated machine learning has emerged as a new methodology to investigate and address related problems. The combined use of machine learning and neurorobotics allows us to solve problems and find explanatory models that would not be possible with traditional techniques, which are basic within the principles of symmetry. Hence, neuro-robotics has become a new research field. Accordingly, this study aimed to classify existing publications on neurorobotics via content analysis and knowledge mapping. The study also aimed to effectively understand the development trend of neurorobotics-integrated machine learning. Based on data collected from the Web of Science, 46 references were obtained, and bibliometric data from 2013 to 2021 were analyzed to identify the most productive countries, universities, authors, journals, and prolific publications in neurorobotics. CiteSpace was used to visualize the analysis based on co-citations, bibliographic coupling, and co-occurrence. The study also used keyword network analysis to discuss the current status of research in this field and determine the primary core topic network based on cluster analysis. Through the compilation and content analysis of specific bibliometric analyses, this study provides a specific explanation for the knowledge structure of the relevant subject area. Finally, the implications and future research context are discussed as references for future research.

A Survey on Design, Actuation, Modeling, and Control of Continuum Robot

In this paper, we describe the advances in the design, actuation, modeling, and control field of continuum robots. After decades of pioneering research, many innovative structural design and actuation methods have arisen. Untethered magnetic robots are a good example; its external actuation characteristic allows for miniaturization, and they have gotten a lot of interest from academics. Furthermore, continuum robots with proprioceptive abilities are also studied. In modeling, modeling approaches based on continuum mechanics and geometric shaping hypothesis have made significant progress after years of research. Geometric exact continuum mechanics yields apparent computing efficiency via discrete modeling when combined with numerical analytic methods such that many effective model-based control methods have been realized. In the control, closed-loop and hybrid control methods offer great accuracy and resilience of motion control when combined with sensor feedback information. On the other hand, the advancement of machine learning has made modeling and control of continuum robots easier. The data-driven modeling technique simplifies modeling and improves anti-interference and generalization abilities. This paper discusses the current development and challenges of continuum robots in the above fields and provides prospects for the future.

Model-Based Nonlinear Feedback Controllers for Pressure Control of Soft Pneumatic Actuators Using On/Off Valves

This article describes the application and comparison of three nonlinear feedback controllers for low-level control of soft actuators driven by a pressure source and single high-speed on/off solenoid valve. First, a mathematical model of the pneumatic system is established and the limitations of the open-loop system are evaluated. Next, a model of the pneumatic system is developed using Simscape Fluids to evaluate the performance of various control strategies. In this article, State-Dependent Riccati Equation control, sliding mode control, and feedback linearization are considered. To improve robustness to model uncertainties, the sliding mode and feedback linearization control strategies are augmented with integral action. The model of the pneumatic system is also used to develop a feedforward component, which is added to a PI controller with anti-windup. The simulation and experimental results demonstrate the effectiveness of the proposed controllers for pressure tracking.

Design of sensing system for experimental modeling of soft actuator applied for finger rehabilitation

Safe interaction and inherent compliance with soft robots have motivated the evolution of soft rehabilitation robots. Among these, soft robotic gloves are known as an effective tool for stroke rehabilitation. This research proposed a pneumatically actuated soft robotic for index finger rehabilitation. The proposed system consists of a soft bending actuator and a sensing system equipped with four inertial measurement unit sensors to generate kinematic data of the index finger. The designed sensing system can estimate the range of motion (ROM) of the finger’s joints by combining angular velocity and acceleration values with the standard Kalman filter. The sensing system is evaluated regarding repeatability and reliability through static and dynamic experiments in the first step. The root mean square error attained in static and dynamic states are 2 $^\circ$ and 3 $^\circ$ , sequentially, representing an efficient function of the fusion algorithm. In the next step, experimental models have been developed to analyze and predict a soft actuator’s behavior in free and constrained states using the sensing system’s data. Thus, parametric system identification methods, artificial neural network—multilayer perceptron (ANN-MLP), and artificial neural network—radial basis function algorithms (ANN-RBF) have been compared to achieve an optimal model. The results reveal that ANN models, particularly RBF ones, can predict the actuator behavior with reasonable accuracy in the free and constrained state (<1 $^\circ$ ). Hence, the need for intricate analytical modeling and material characterization will be eliminated, and controlling the soft actuator will be more practical. Besides, it assesses the ROM and finger functionality.

A Survey for Machine Learning-Based Control of Continuum Robots

Soft continuum robots have been accepted as a promising category of biomedical robots, accredited to the robots’ inherent compliance that makes them safely interact with their surroundings. In its application of minimally invasive surgery, such a continuum concept shares the same view of robotization for conventional endoscopy/laparoscopy. Different from rigid-link robots with accurate analytical kinematics/dynamics, soft robots encounter modeling uncertainties due to intrinsic and extrinsic factors, which would deteriorate the model-based control performances. However, the trade-off between flexibility and controllability of soft manipulators may not be readily optimized but would be demanded for specific kinds of modeling approaches. To this end, data-driven modeling strategies making use of machine learning algorithms would be an encouraging way out for the control of soft continuum robots. In this article, we attempt to overview the current state of kinematic/dynamic model-free control schemes for continuum manipulators, particularly by learning-based means, and discuss their similarities and differences. Perspectives and trends in the development of new control methods are also investigated through the review of existing limitations and challenges.

Robust Fractional-Order Control Using a Decoupled Pitch and Roll Actuation Strategy for the I-Support Soft Robot

Tip control is a current open issue in soft robotics; therefore, it has received a good amount of attention in recent years. The desirable soft characteristics of these robots turn a well-solved problem in classic robotics, like the end-effector kinematics and dynamics, into a challenging problem. The high redundancy condition of these robots hinders classical solutions, resulting in controllers with very high computational costs. In this paper, a simplification is proposed in the actuation setup of the I-Support soft robot, allowing the use of simple strategies for tip inclination control. In order to verify the proposed approach, inclination step input and trajectory-tracking experiments were performed on a single module of the I-Support robot, resulting in zero output error in all cases, including those where the system was exposed to disturbances. The comparative results of the proposed controllers, a proportional integral derivative (PID) and a fractional order robust (FOPI) controller, validate the feasibility of the proposed approach, showing a clear advantage in the use of the fractional robust controller for the tip inclination control of the I-Support robot compared to the integer order controller.

A Recurrent Neural-Network-Based Real-Time Dynamic Model for Soft Continuum Manipulators

This paper introduces and validates a real-time dynamic predictive model based on a neural network approach for soft continuum manipulators. The presented model provides a real-time prediction framework using neural-network-based strategies and continuum mechanics principles. A time-space integration scheme is employed to discretize the continuous dynamics and decouple the dynamic equations for translation and rotation for each node of a soft continuum manipulator. Then the resulting architecture is used to develop distributed prediction algorithms using recurrent neural networks. The proposed RNN-based parallel predictive scheme does not rely on computationally intensive algorithms; therefore, it is useful in real-time applications. Furthermore, simulations are shown to illustrate the approach performance on soft continuum elastica, and the approach is also validated through an experiment on a magnetically-actuated soft continuum manipulator. The results demonstrate that the presented model can outperform classical modeling approaches such as the Cosserat rod model while also shows possibilities for being used in practice.

Heterogeneous sensing in a multifunctional soft sensor for human-robot interfaces

Soft sensors have been playing a crucial role in detecting different types of physical stimuli to part or the entire body of a robot, analogous to mechanoreceptors or proprioceptors in biology. Most of the currently available soft sensors with compact form factors can detect only a single deformation mode at a time due to the limitation in combining multiple sensing mechanisms in a limited space. However, realizing multiple modalities in a soft sensor without increasing its original form factor is beneficial, because even a single input stimulus to a robot may induce a combination of multiple modes of deformation. Here, we report a multifunctional soft sensor capable of decoupling combined deformation modes of stretching, bending, and compression, as well as detecting individual deformation modes, in a compact form factor. The key enabling design feature of the proposed sensor is a combination of heterogeneous sensing mechanisms: optical, microfluidic, and piezoresistive sensing. We characterize the performance on both detection and decoupling of deformation modes, by implementing both a simple algorithm of threshold evaluation and a machine learning technique based on an artificial neural network. The proposed soft sensor is able to estimate eight different deformation modes with accuracies higher than 95%. We lastly demonstrate the potential of the proposed sensor as a method of human-robot interfaces with several application examples highlighting its multifunctionality.

Novel Design and Modeling of a Soft Pneumatic Actuator Based on Antagonism Mechanism

The soft actuator possesses the characteristics of flexibility, environmental adaptability, and human–machine interaction. Firstly, aiming to resolve the limitation of variable stiffness performance of a traditional pneumatic artificial muscle (PAM) actuator, based on the antagonistic mechanism of extensor and contractor muscles, a novel pneumatic soft actuator coupled of extensor and contractor muscles is proposed in this paper. The actuator can perform the compound action of elongation/contraction, and the stiffness of it can be controlled by adjusting the elongation and contraction forces. Secondly, based on the deformation principle of woven and elastic fabric layers, the mechanical characteristics model of the actuator is established and simulated. The mechanical properties of the actuator are tested under different pressures and deformation displacement and the variable stiffness characteristics of the actuator are verified. Finally, actuators are utilized to manufacture a soft mechanical manipulator, which can achieve variable stiffness in a fixed bending attitude.

Biomedical soft robots: current status and perspective

This paper reviews the current status of soft robots in biomedical field. Soft robots are made of materials that have comparable modulus of elasticity to that of biological systems. Several advantages of soft robots over rigid robots are safe human interaction, ease of adaptation with wearable electronics and simpler gripping. We review design factors of soft robots including modeling, controls, actuation, fabrication and application, as well as their limitations and future work. For modeling, we survey kinematic, multibody and numerical finite element methods. Finite element methods are better suited for the analysis of soft robots, since they can accurately model nonlinearities in geometry and materials. However, their real-time integration with controls is challenging. We categorize the controls of soft robots as model-based and model-free. Model-free controllers do not rely on an explicit analytical or numerical model of the soft robot to perform actuation. Actuation is the ability to exert a force using actuators such as shape memory alloys, fluid gels, elastomers and piezoelectrics. Nonlinear geometry and materials of soft robots restrict using conventional rigid body controls. The fabrication techniques used for soft robots differ significantly from that of rigid robots. We survey a wide range of techniques used for fabrication of soft robots from simple molding to more advanced additive manufacturing methods such as 3D printing. We discuss the applications and limitations of biomedical soft robots covering aspects such as functionality, ease of use and cost. The paper concludes with the future discoveries in the emerging field of soft robots.

Soft Miniaturized Actuation and Sensing Units for Dynamic Force Control of Cardiac Ablation Catheters

Recently, there has been active research in finding robotized solutions for the treatment of atrial fibrillation (AF) by augmenting catheter systems through the integration of force sensors at the tip. However, limited research has been aimed at providing automatic force control by also integrating actuation of the catheter tip, which can significantly enhance safety in such procedures. This article solves the demanding challenge of miniaturizing both actuation and sensing for integration into flexible catheters. Fabrication strategies are presented for a series of novel soft thick-walled cylindrical actuators, with embedded sensing using eutectic gallium-indium. The functional catheter tips have a diameter in the range of 2.6-3.6 mm and can both generate and detect forces in the range of < 0.4 N, with a bandwidth of 1-2 Hz. The deformation modeling of thick-walled cylinders with fiber reinforcement is presented in the article. An experimental setup developed for static and dynamic characterization of these units is presented. The prototyped units were validated with respect to the design specifications. The preliminary force control results indicate that these units can be used in tracking and control of contact force, which has the potential to make AF procedures much safer and more accurate.

Research on abnormal object detection in specific region based on Mask R-CNN

As an information carrier with rich semantics, image plays an increasingly important role in real-time monitoring of logistics management. Abnormal objects are typically closely related to the specific region. Detecting abnormal objects in the specific region is conducive to improving the accuracy of detection and analysis, thereby improving the level of logistics management. Motivated by these observations, we design the method called abnormal object detection in a specific region based on Mask R-convolutional neural network: Abnormal Object Detection in Specific Region. In this method, the initial instance segmentation model is obtained by the traditional Mask R-convolutional neural network method, then the region overlap of the specific region is calculated and the overlapping ratio of each instance is determined, and these two parts of information are fused to predict the exceptional object. Finally, the abnormal object is restored and detected in the original image. Experimental results demonstrate that our proposed Abnormal Object Detection in Specific Region can effectively identify abnormal objects in a specific region and significantly outperforms the state-of-the-art methods.

A review of recent advancements in soft and flexible robots for medical applications

BACKGROUND

Soft and flexible robots for medical applications are needed to change their flexibility over a wide range to perform tasks adequately. The mechanism and theory of flexibility has been a scientific issue and is of interest to the community.

METHODS

Recent advancements of bionics, flexible actuation, sensing, and intelligent control algorithms as well as tunable stiffness have been referenced when soft and flexible robots are developed. The benefits and limitations of these relevant studies and how they affect the flexibility are discussed, and possible research directions are explored.

RESULTS

The bionic materials and structures that demonstrate the potential capabilities of the soft medical robot flexibility are the fundamental guarantee for clinical medical applications. Flexible actuation that used to provide power, intelligent control algorithms which are the exact executors, and the wide range stiffness of the soft materials are the three important influence factors for soft medical robots.

CONCLUSION

Some reasonable suggestions and possible solutions for soft and flexible medical robots are proposed, including novel materials, flexible actuation concepts with a built-in source of energy or power, programmable flexibility, and adjustable stiffness.

3D-Printing and Machine Learning Control of Soft Ionic Polymer-Metal Composite Actuators

This paper presents a new manufacturing and control paradigm for developing soft ionic polymer-metal composite (IPMC) actuators for soft robotics applications. First, an additive manufacturing method that exploits the fused-filament (3D printing) process is described to overcome challenges with existing methods of creating custom-shaped IPMC actuators. By working with ionomeric precursor material, the 3D-printing process enables the creation of 3D monolithic IPMC devices where ultimately integrated sensors and actuators can be achieved. Second, Bayesian optimization is used as a learning-based control approach to help mitigate complex time-varying dynamic effects in 3D-printed actuators. This approach overcomes the challenges with existing methods where complex models or continuous sensor feedback are needed. The manufacturing and control paradigm is applied to create and control the behavior of example actuators, and subsequently the actuator components are combined to create an example modular reconfigurable IPMC soft crawling robot to demonstrate feasibility. Two hypotheses related to the effectiveness of the machine-learning process are tested. Results show enhancement of actuator performance through machine learning, and the proof-of-concepts can be leveraged for continued advancement of more complex IPMC devices. Emerging challenges are also highlighted.

Design of robot finger based on flexible tactile sensor

In this article, a flexible tactile sensor that made of conductive silicone rubber for dexterous robot hand is designed. The tactile sensor is made up of four microsensors. The maximum gripping force is simulated when the degree of a robot finger joint is 138. Meanwhile, a control system to analyze the creep and hysteresis characteristics and a processing system of the tactile sensor is designed. We also demonstrated an experiment for the application of robot grasp object, showing the finger’s flexibility and sensitivity. Then the feedback data is sent to control system to provide precise grasp action changes for the robot hand.

Design and Preliminary Testing of a Continuum Assistive Robotic Manipulator

Background: The application of continuum manipulators as assistive robots is discussed and tested through the use of Bendy ARM, a simple manually teleoperated tendon driven continuum manipulator prototype. Methods: Two rounds of user testing were performed to evaluate the potential of this arm to aid people living with disabilities in completing activities of daily living. Results: In the first round of user testing, 14 able-bodied subjects successfully completed the prescribed task (pick-and-place) using multiple control schemes after being given a brief introduction and one minute of practice with each scheme. In the second round of user testing, subjects ( n = 3 ) demonstrated between 29.5 and 48.9 percent improvement in completion time across twelve trials of a peg-in-hole task, and between 8.4 and 33.8 percent improvement across six trials of a task involving opening and closing a drawer. Conclusion: Based on these results, it is posited that continuum manipulators merit further consideration as a safer and more cost-effective alternative to existing commercially available assistive robotic manipulators.

A Self-Deformation Robot Design Incorporating Bending-Type Pneumatic Artificial Muscles

With robots becoming closer to humans in recent years, human-friendly robots made of soft materials provide a new line of research interests. We designed and developed a soft robot that can move via self-deformation toward the practical application of monitoring children and the elderly on a daily basis. The robot’s structure was built out of flexible frames, which are bending-type pneumatic artificial muscles (BPAMs). We first provide a description and discussion on the nature of BPAM, followed by static characteristics experiment. Although the BPAM theoretical model shares a similar tendency with the experimental results, the actual BPAMs moved along the depth direction. We then proposed and demonstrated an effective locomotion method for the robot and calculated its locomotion speed by measuring its drive time and movement distance. Our results confirmed the reasonability of the robot’s speed for monitoring children and the elderly. Nevertheless, during the demonstration, some BPAMs were bent sharply by other activated BPAMs as the robot was driving, leaving a little damage on these BPAMs. This will be addressed in our future work.

Soft Poly-Limbs: Toward a New Paradigm of Mobile Manipulation for Daily Living Tasks

We present the design and development of the fluid-driven, wearable, Soft Poly-Limb (SPL), from the Greek word polys, meaning many. The SPL utilizes the numerous traits of soft robotics to enable a novel approach in providing safe and compliant mobile manipulation assistance to healthy and impaired users. This wearable system equips the user with a controllable additional limb that is capable of complex three-dimensional motion in space. Similar to an elephant trunk, the SPL is able to manipulate objects using a variety of end effectors, such as suction adhesion or a soft grasper, as well as its entire soft body to conform around an object, able to lift 2.35 times its own weight. To develop these highly articulated soft robotic limbs, we provide a novel set of systematic design rules, obtained through varying geometrical parameters of the SPL through experimentally verified finite element method models. We investigate performance of the limb by testing the lifetime of the new SPL actuators, evaluating its payload capacity, operational workspace, and capability of interacting close to a user through a spatial mobility test. Furthermore, we are able to demonstrate limb controllability through multiple user-intent detection modalities. Finally, we explore the limb's ability to assist in multitasking and pick and place scenarios with varying mounting locations of the SPL around the user's body. Our results highlight the SPL's ability to safely interact with the user while demonstrating promising performance in assisting with a wide variety of tasks, in both work and general living settings.