Flexible wearable devices have attracted wide attention in capacious fields because of their real-time and continuous monitoring of human information. The development of flexible sensors and corresponding integration with wearable devices is of great significance to build smart wearable devices. In this work, multi-walled carbon nanotube/polydimethylsiloxane-based (MWCNT/PDMS) resistive strain sensors and pressure sensors were developed to integrate a smart glove for human motion/perception detection. Firstly, MWCNT/PDMS conductive layers with excellent electrical and mechanical properties (resistivity of 2.897 KΩ · cm, elongation at break of 145%) were fabricated via a facile scraping-coating method. Then, a resistive strain sensor with a stable homogeneous structure was developed due to the similar physicochemical properties of the PDMS encapsulation layer and MWCNT/PDMS sensing layer. The resistance changes of the prepared strain sensor exhibited a great linear relationship with the strain. Moreover, it could output obvious repeatable dynamic response signals. It still had good cyclic stability and durability after 180° bending/restoring cycles and 40% stretching/releasing cycles. Secondly, MWCNT/PDMS layers with bioinspired spinous microstructures were formed by a simple sandpaper retransfer process and then assembled face-to-face into a resistive pressure sensor. The pressure sensor presented a linear relationship of relative resistance change and pressure in the range of 0–31.83 KPa with a sensitivity of 0.026 KPa⁻¹, and a sensitivity of 2.769 × 10⁻⁴ KPa⁻¹ over 32 KPa. Furthermore, it responded quickly and kept good cycle stability at 25.78 KPa dynamic loop over 2000 s. Finally, as parts of a wearable device, resistive strain sensors and a pressure sensor were then integrated into different areas of the glove. The cost-effective, multi-functional smart glove can recognize finger bending, gestures, and external mechanical stimuli, which holds great potential in the fields of medical healthcare, human-computer cooperation, and so on.

Hand dexterity is tremendously valuable to robots for task-dependent manipulation and interacting with the world. In this work, we present a novel soft pneumatic dexterous hand, which demonstrates highly dexterous and versatile anthropomorphic properties. Inspired by human hand, the proposed hand possesses 21 degrees of freedoms (DOFs), is implemented in anthropomorphic fingers, soft parallel palm and the ability to spread the fingers with a modular design. The anthropomorphic fingers are designed in accordance with the joint distribution of the human fingers. By using a layered design, each independent actuator is divided into an actuating layer and a contact layer, which allows the grasping mode and the contact mode to be adjusted in real-time. To streamline the creation of the palm and improve the dexterity of the hand, the soft parallel palm is actuated by bellow actuators, designed by using the characteristics of a compact structure and stable movement of the parallel mechanism. Additionally, the bellow actuators are also used to realize the ability to spread the fingers and assist the carpometacarpal (CMC) joint bending of the thumb through the passive deformation of the flexible joint. Through series of experiments, we evaluate the hand by demonstrating its ability to achieve the highest possible score in the Kapandji test for thumb dexterity, to realize all 33 grasp types of the comprehensive GRASP taxonomy, and to interact with humans in the real world.

The existing soft crawling robots usually have single-motion mode, which results in poor motion adaptability and significantly restricts the application field of the soft crawling robots. To further increase the motion adaptability of the soft crawling robots and expand their application space, in this letter, inspired by the movement mechanism of the snake scales, we present a pneumatic soft crawling robot with multi-modal locomotion based on a deployable and foldable (D-F) mechanism. The robot adopts a four-chamber bellows actuator to further enhance the robot's movement adaptability in motion space. The head and tail parts of the robot are provided with a D-F mechanism imitating snake scales, which can be opened and closed through the extension and contraction of the bellows actuator. By adopting different actuation strategies and utilizing the anisotropy of friction and the alteration of the center of gravity, the robot can achieve multi-modal locomotion such as straight walking, steering, rolling and obstacle crossing to satisfy the requirements of varied surroundings. An open-loop control system of the robot is developed to assess its motion performance. The mathematical model of the robot is established using cosserat-rod theory to clarify the relationship between actuating pressure and robot's motion variables under various motion strategy combinations. The experimental results show that the soft crawling robot designed in this study can efficiently adapt to moving surfaces with different material properties and achieve diverse motion modes under different actuation strategies. Additionally, it can adapt to exploratory tasks in unstructured environment.

To address the zero-shot fault diagnosis challenge in electro-hydrostatic actuator (EHA), a Diffusion-Enhanced Dual-Domain Adversarial Network (DEDDAN) is proposed. Acquired experimental signals were stacked and reorganized in the time domain into grayscale images. A denoising diffusion probabilistic model (DDPM) was then employed to augment the simulation image dataset generated from an EHA simulation model following identical preprocessing. The training set for DEDDAN was constructed by combining experimental normal-condition grayscale images with the augmented simulation dataset. Two domain discriminators were designed to focus on global and local features, while weighting of loss functions balanced training objectives. Experimental results show DEDDAN achieves 95.71% accuracy, outperforming purely data-driven and single-discriminator adversarial methods.

Inspired by identifying directions through the geomagnetic field for migrating birds, in this work, we proposed and fabricated a single-joint worm-like robot with a centimeter scale, the motion of which could be easily guided by a magnet. The robot consists of a pneumatic deformable bellow and a permanent magnet fixed in the bellow’s head that will generate magnetic force and friction. Firstly, in order to clarify the actuating mechanism, we derived the relationship between the elongation of the bellows and the air pressure through the Yeoh constitutive model, which was utilized to optimize the structural parameters of the bellow. Then the casting method is introduced to fabricate the silicone bellow with a size of 20 mm in diameter and 28 mm in length. The manufacturing error of the bellow was evaluated by 3D laser scanning technology. Thereafter, the robot’s moving posture was analyzed by considering the force and corresponding motion state, and the analysis model was established by mechanics theory. The experimental results show that the worm-like robot’s maximum speed can reach 9.6 mm/s on the cardboard. Meanwhile, it exhibits excellent environmental adaptability that can move in pipelines with a diameter of 21 mm, 32 mm, 40 mm, and 50 mm, and surfaces with different roughness. Moreover, the robot’s motion was successfully guided under the presence of the magnetic field, which shows great potential for pipeline detection applications.