Publications

No abstract is provided for this article.

Humans can accomplish complex contact-rich tasks using vision and touch, with highly reactive capabilities such as fast response to external changes and adaptive control of contact forces; however, this remains challenging for robots. Existing visual imitation learning (IL) approaches rely on action chunking to model complex behaviors, which lacks the ability to respond instantly to real-time tactile feedback during the chunk execution. Furthermore, most teleoperation systems struggle to provide fine-grained tactile / force feedback, which limits the range of tasks that can be performed. To address these challenges, we introduce TactAR, a low-cost teleoperation system that provides real-time tactile feedback through Augmented Reality (AR), along with Reactive Diffusion Policy (RDP), a novel slow-fast visual-tactile imitation learning algorithm for learning contact-rich manipulation skills. RDP employs a two-level hierarchy: (1) a slow latent diffusion policy for predicting high-level action chunks in latent space at low frequency, (2) a fast asymmetric tokenizer for closed-loop tactile feedback control at high frequency. This design enables both complex trajectory modeling and quick reactive behavior within a unified framework. Through extensive evaluation across three challenging contact-rich tasks, RDP significantly improves performance compared to state-of-the-art visual IL baselines. Furthermore, experiments show that RDP is applicable across different tactile / force sensors. Code and videos are available on https://reactive-diffusion-policy.github.io.

ABSTRACT Millimeter-scale soft robots (milli-SRs) promise significant advancements in biomedical engineering and inspection, enabling precise navigation in confined spaces. However, fabricating miniaturized fluid-driven soft robots is hindered by microscale forces. Here, we introduce a new universal design and fabrication approach (referred to as the mini bubble casting method) to create high-quality multifunctional fluid-driven milli-SRs. By injecting a bubble into pre-modified silicone liquid under high-stability conditions, we achieve submillimeter internal voids, overcoming interfacial-tension-induced instability. The modification strategy is guided by our theoretical model, which explains the influence of viscous resistance and interfacial tension on the dynamic behavior of the bubble-silicone interface. We successfully fabricate soft milli-actuators ten times smaller than existing works with low surface smoothness (${\rm Ra}=11.2$ nm). We demonstrate a milli-gripper handling delicate insects and a thrombus extractor for narrow vessels. We present a miniature steerable tip for bronchial navigation, improving safety and dexterity over traditional tools, showing the tremendous biomedical potential of these devices.

A slim manipulator with redundant degrees of freedom can be of great use for inspection and maintenance in confined spaces. Cable-actuation is often preferred because of the possible proximal placement of the actuators. However, tension induced elongation and friction of the cables may substantially vary under different manipulator poses. It is then difficult to achieve high positioning accuracy under different loading conditions, even when such a manipulator went through careful motion calibration and actuation compensation. This paper hence proposes a 12-section cable-driven hyper-redundant manipulator with 24 degrees of freedom and closed-loop control. Customized magnetic angle sensors were integrated at each joint with a carefully designed sensor communication network to achieve rapid joint feedback. A novel puller-follower controller was proposed to properly actuate three cables for each universal joint that connects adjacent manipulator sections. The two joint angles of the universal joint and the cable tensions can be simultaneously controlled with the proposed puller-follower controller. The design overview, system descriptions, kinematics and experimental characterizations are reported in detail. After automatic straightening from arbitrary initial statuses, the proposed manipulator can achieve ±0.5° precision at each joint, even under a tip load up to 500 g.

Electrically driven dielectric elastomer artificial muscles represent a transformative advancement in the field of soft robotics. However, their output performance has encountered a bottleneck owing to the insufficient electromechanical sensitivity of dielectric elastomers. We present a hetero-cross-linking–induced phase separation strategy to design semiseparated biphasic bicontinuous dielectric elastomers with a high electromechanical sensitivity of 360 per megapascal. Our strategy harnesses varying silicone elastomer cross-linking mechanisms to form an interconnected dielectric phase within a soft mechanical phase in the resultant elastomers. These elastomer-based artificial muscles simultaneously exhibit high energy density and power density, as well as ultralong life span under low-driving fields. Applications involve a robotic arm with large stroke and untethered soft crawling robots with multimodal locomotion, showcasing their versatility.

Ionogels have attracted considerable attention as versatile materials due to their unique ionic conductivity and thermal stability. However, relatively weak mechanical performance of many existing ionogels has hindered their broader application. Herein, we develop robust, tough, and impact‐resistant mechanically interlocked network ionogels (IGMINs) by incorporating ion liquids with mechanical bonds that can dissipate energy while maintain structural stability. Profiting from the dynamic yet stable nature of the mechanically interlocked networks, IGMINs exhibit high tensile strength (9.6 MPa), fracture energy (39 kJ/m2), and toughness (25.9 MJ/m3), along with a high elongation rate (473%) and excellent impact resistance and shape memory, resulting in overall performance that surpasses most reported ionogels. Furthermore, in the application of strain sensors for monitoring the gait of crawling robots, the toughness and robustness of IGMINs ensure their ability to consistently output stable electrical signals during the stretching and contraction processes, thereby highlighting their practical application potential. Our work provides a new research strategy for toughening ionogels and promotes the development of mechanically interlocked materials.

A homogeneous pneumatic soft robot may generate complex output motions using a simple input pressure, resulting from its morphological shape that locally deforms the soft material to different degrees by simultaneously tailoring the structural characteristics and orienting the input pressure. To date, design of the morphological shape (inverse problem) has not been fully addressed. This article outlines a geometry–mechanics–optimization integrated approach to automatically shaping a pneumatic soft actuator or robot that achieves the desired deformation behavior. Instead of constraining the robot's geometry within any predefined regular shape, we employ B-splines to allow generation of freeform boundary surfaces, and use nonlinear mechanical modelling and shape derivative based optimization to navigate the high-dimensional design space. Our design framework can readily regulate the surface quality during the morphological evolution, by imposing the geometric constraints in terms of the principal curvatures and the minimal distance between surfaces as penalty functions. The effect of external forces including the gravity and the interaction force at the end-effector is also taken into account to generalize the method for design problems in which the load capability is also pursued. To improve the computational efficiency, suboptimization problems are constructed within a trust region in which the displacement-dependent objective function is approximated by its first-order Taylor polynomial based on the gradient information to avoid frequently performing time-consuming nonlinear finite element analysis. The suboptimization problems are then solved by the quasi-Newton method combined with the backtracking line search strategy. We showcase various applications to validate our design approach, including actuators for basic extension, bending, and twisting motions, and continuous robot arms that can perform desired in-plane and out-of-plane configurations. We also show that our method can address design of multiple chambers for achieving multiple target deformation behaviors, by co-optimizing the morphological shape and air pressures, which is validated by two examples.

larvae can quickly ascend tubular surfaces via rolling, a capability not yet replicated by soft robots. Here, we present a single-piece soft robot capable of rolling along tubular structures by sequentially actuating its built-in axial muscles. We reveal that the sequential actuation generates distributed spinning torques along the robot's curved axis, enabling continuous non-coaxial rolling-distinct from current gravity-dependent rolling solutions. This non-coaxial rolling mechanism allows the robot to swiftly navigate tubular surfaces while conforming to their shapes and maintaining a stable grip. The robot's deformation and gripping force are actively adjusted to enhance its adaptability to various surfaces. We demonstrate that our robot can ascend pipes with varying geometries (e.g., varying-diameter, spiral-shaped, or non-cylindrical), traverse diverse terrains, pass through confined tunnels, and transition smoothly between planar rolling and pipe climbing. The robot's great adaptability and rapid movement underscore its potential for navigating scenarios with intricate surface geometries.

Multi-chamber soft pneumatic actuators (m-SPAs) have been widely used in soft robotic systems to achieve versatile grasping and locomotion. However, existing m-SPAs have slow actuation speed and are either limited by a finite air supply or require energy-consuming hardware to continuously supply compressed air. Here, we address these shortcomings by introducing an internal exhaust air recirculation (IEAR) mechanism for high-speed and low-energy actuation of m-SPAs. This mechanism recirculates the exhaust compressed air and recovers the energy by harnessing the rhythmic actuation of multiple chambers. We develop a theoretical model to guide the analysis of the IEAR mechanism, which agrees well with the experimental results. Comparative experimental results of several sets of m-SPAs show that our IEAR mechanism significantly improves the actuation speed by more than 82.4% and reduces the energy consumption per cycle by more than 47.7% under typical conditions. We further demonstrate the promising applications of the IEAR mechanism in various pneumatic soft machines and robots such as a robotic fin, fabric-based finger, and quadruped robot. Corresponding author(s) Email: guguoying@sjtu.edu.cn