Soft robotics is a rapidly evolving field where robots are fabricated using highly deformable materials and usually follow a bioinspired design. Their high dexterity and safety makes them ideal for applications such as gripping, locomotion, and biomedical devices, where the environment is highly dynamic and sensitive to physical interaction. Pneumatic actuation remains the dominant technology in soft robotics due to its low cost and mass, fast response time, and easy implementation. Given the significant number of publications in soft robotics over recent years, newcomers and even established researchers may have difficulty assessing the state of the art. To address this issue, this article summarizes the development of soft pneumatic actuators and robots up until the date of publication. The scope of this article includes the design, modeling, fabrication, actuation, characterization, sensing, control, and applications of soft robotic devices. In addition to a historical overview, there is a special emphasis on recent advances such as novel designs, differential simulators, analytical and numerical modeling methods, topology optimization, data-driven modeling and control methods, hardware control boards, and nonlinear estimation and control techniques. Finally, the capabilities and limitations of soft pneumatic actuators and robots are discussed and directions for future research are identified.

Soft-bodied robotic manipulators have great potential for use in minimally invasive surgery, owing to their advantages of high flexibility with infinite degrees of freedom (DOF). One of the potential applications is to perform blood suctioning, which is inevitable during the surgery. To attain higher efficiency in suctioning, the robotic tip should remain vertical while moving along on the work surface. Motivated by this application, this article presents a novel soft robot design and its control scheme to properly configure the tip of a two-segment soft robot while following a planned trajectory on the work surface. Aiming to reduce the incision size and the possibility of infection, a 3D-printed soft-bodied manipulator utilizing the cable-driven mechanism with a diameter of 9 mm was designed and fabricated. An additional DOF was added through a motorized insertion stage. The robot system was modeled using piecewise constant-curvature assumption, and an RGB-D vision was employed to enhance the accuracy of the kinematic-based controller. Performances of the tip positioning and verticalizing were evaluated via simulation, and further verified through experiments. The results confirm that the manipulator is capable of following different trajectories at various velocities while keeping its tip vertical. Compared to other similar works, our results are satisfactory with an root-mean-square error of trajectory tracking within 7 mm, and a maximum angular deviation of 6 $^\circ$ . Fluid suction experiments were conducted to demonstrate its effectiveness for automated 3-D suction. This article offers a new tool to support the surgeons for surgical blood suction.

A cable-driven soft robot with redundancy can perform the tip trajectory tracking task and in the meanwhile fulfill some extra constraints, such as tracking with a designated tip orientation, or avoiding obstacles in the environment. These constraints require proper motion planning of the soft material-based body that can be axially compressed. In this letter, we derive the compressible curvature kinematics of a cable-driven soft robot which takes the undesirable axial compression caused by the cable-driven mechanism into account. Robot motion planning for tip trajectory tracking tasks in constrained conditions, including fixed orientation tip and manipulator–obstacle collision avoidance, are investigated. The inverse solution of cable actuation is formulated as a damped least-square optimization problem and iteratively computed off-line. The performance of path and trajectory tracking, and the obedience to constraints are evaluated via the simulation we make open-source, as well as prototype experiments. The method can be generalized to similar multisegment cable-driven soft robotic systems by customizing the robot parameters for the prior motion planning in tip trajectory following tasks.

Finding the shape of a continuum robot is a challenging task, especially in a constrained environment where external sensors like cameras are unable to provide sufficient information. In particular, accurate robot control becomes more difficult when the robot needs to interact with uncertain external payloads. In this article, we present the development of a soft robotic system with embedded sensors to help reconstruct the shape of the robot under different external disturbances. First, strain gauges were employed to perceive the current deformation state of the robot due to actuator inputs and external payloads. Then, a shape reconstruction (SR) method based on a spatial curve fitting approach was proposed, where multiple control points along the robot were predicted using Neural Networks. With the estimated shape, a local inverse kinematics model was developed so that the robot end-effector can be precisely controlled in quasi-static state to reach various positions, and position of end effector calculated via SR module was employed to improve the path following performance. Simulation and experiments were conducted and the results confirm that the proposed control system is effective to predict the shape and properly control the robot with or without external payloads.

A soft-bodied robot exhibits prominent dexterity due to the soft nature of its material. However, the softness can become a burden when the robot needs to interact with the environment, given that the targeted object is usually much stiffer than the compliant soft robot. A variable-stiffness soft robot, fusing the merits of softness and stiffness, is in favor of many applications, such as robot-assisted minimally invasive surgeries. In this article, we propose a tendon-tensioning method to adaptively control the stiffness of a dual-segment tendon-driven backboneless soft robot based on depth vision. A depth-vision-based closed-loop controller is designed for stiffness compensation when the manipulator is subjected to the external load. Experiments were conducted to examine the feasibility and performance of the proposed method. The results confirm our control scheme on the robot with controllability of stiffness up to 132%. Based on our method, the manipulator with an external payload can follow designated trajectories with positioning errors reduced up to 50% comparing to that with open-loop control. Without quantifying the instantaneous stiffness, this work contributes a generalized method for tuning the stiffness of the tendon-driven soft robots in the presence of external disturbances without onboard sensing.

To steer a soft robot precisely in an unconstructed environment with minimal collision remains an open challenge for soft robots. When the environments are unknown, prior motion planning for navigation may not always be available. This article presents a novel Sim-to-Real method to guide a cable-driven soft robot in a static environment under the Simulation Open Framework Architecture (SOFA). The scenario aims to resemble one of the steps during a simplified transoral tracheal intubation process where a robotic endotracheal tube is guided to the upper trachea–larynx location by a flexible video-assisted endoscope/stylet. In SOFA, we employ the quadratic programming inverse solver to obtain collision-free motion strategies for the endoscope/stylet manipulation based on the robot model and encode the virtual eye-in-hand vision. Then, we associate the anatomical features recognized by the virtual vision and the joint space motion using a closed-loop nonlinear autoregressive exogenous model (NARX) network. Afterward, we transfer the learned knowledge to the robot prototype, expecting it to navigate to the desired spot in a new phantom environment automatically based on its eye-in-hand vision only. Experiment results indicate that our soft robot can efficaciously navigate through the unstructured phantom to the desired spot with minimal collision motion according to what it has learned from the virtual environment. The results show that the average R-squared coefficient between the closed-loop NARX-forecasted and SOFA-referenced robot's cable and prismatic joint space motion are 0.963 and 0.997, respectively. The eye-in-hand visions also demonstrate a good alignment between the robot tip and the glottis.