Tensegrity mobile robots, known for robust terrain adaptability and impressive stiffness-to-mass ratio, have become a focal point in research. Particularly, the innovative design of multilocomotion tensegrity robots is challenging due to their complex parameters and flexible components. Achieving the design of a multilocomotion tensegrity mobile robot, which encompasses both topology and dimension, remains a significant challenge. Here, we introduce a design approach combining machine learning technology and incremental model. Utilizing our approach, the redundant design parameters (such as topology, dimension, stiffness of elastic cable, and driving laws under specific terrains) of the tensegrity robot could be acquired without complex mathematical models and intricate iterative design processes. The proposed innovative design approach yields an approximately 12.47-fold increase in efficiency for obtaining the first feasible tensegrity robot configuration compared to enumeration, thereby significantly reducing development time and improving overall design efficiency. Furthermore, our machine learning-driven design approach has demonstrated robust generalization capabilities, facilitating the formulation of driving laws for navigating unknown terrains, thus paving new ways for mobile robots.

Continuum robots are highly suitable for interactive operations in confined spaces due to their inherent body compliance. However, traditional continuum robots with only bending degrees of freedom (DOFs) are limited in both workspace and dexterity, especially when manipulating an object near the proximal part of the arm. This article introduces a deployable continuum robot with a coiling mechanism, DCR, capable of fully stowing, or deploying, its arm into, or out of, a robot case with the help of a coiling mechanism. Deploying the continuum arm provides the robot with variable kinematics and workspace. A variable kinematics motion control framework is then established to enable the end-effector of the robot to smoothly travel through whole workspace while maintaining an optimal configuration. Based on this framework, visual servoing control methods can be applied to the robotic system, achieving high positioning precision up to 1 mm. Experimental results demonstrate that the proposed DCR enlarge the workspace and improve the manipulating dexterity, which are particularly beneficial to applications in confined spaces.

Tensegrity structures offer unique advantages in designing mobile robots with lightweight property, high compliance, and superior stiffness-to-mass ratio. However, the limited motion modes and poor adaptability to diverse terrains of current tensegrity mobile robots significantly hinder their practical applications. For making up the disadvantages, this study presents a novel multi-locomotion tensegrity mobile robot based on a modular design idea. Unlike other tensegrity robots, the tensegrity robot proposed in this work is capable of performing seven distinct locomotion modes and can adapt to seven typical terrains. This study also makes significant progress in establishing the generalized kinematic modeling framework for the multi-locomotion tensegrity robot, enabling the systematic design of locomotion strategies across various terrains. The results of multi-locomotion design show that the proposed robot can adapt to a variety of terrains including flat ground, slope, confined space, obstacle, gap, rough terrain and soft terrain under seven distinct motion modes in terms of earthworm-like, inchworm-like, lateral, lifting, winding, rolling and hybrid locomotion, significantly overcoming the terrain adaptability limitations of existing tensegrity mobile robots. Furthermore, a seamless mode-switching mechanism is proposed for continuous terrain transitions. A prototype with seven modules is constructed to experimentally verify the robot’s multi-locomotion capability and excellent terrain adaptability. The results confirm the robot’s potential for real-world applications in complex environments, offering a robust solution for future mobile robotics challenges.

Patients with lower limb motor dysfunction typically require supine rehabilitation during early recovery. Compared to rigid-structure robots, cable-driven supine rehabilitation robots offer distinct safety and cost advantages. However, existing such robots are generally limited to passive training, particularly those with under-constrained cable-suspended structures. This study proposes a novel admittance model for the under-constrained cable-suspended system for active training in supine status, providing resistance along a specified trajectory. To enhance the training experience, two tunnel models orthogonal to the trajectory were proposed, rigid mode and compliant mode. They meet personalized rehabilitation requirements better by adjusting the impedance model parameters. To prove the efficiency of the under-constrained active training method, a cable-driven parallel robot (CDPR) for lower limb supine training was developed, achieving multi-joint lower limb movement via four cables. Considering that positioning of patient relative to the proposed robot is necessary for performing active training, an autonomous patient position identification method is introduced, without extra sensors. Experiments on the prototype confirmed the position identification method's feasibility and the under-constrained active training control model's effectiveness in both tunnel modes.

In mobile robots, wheeled-legged robots are propelled by C-shaped legs that rotate around the circumference. They combine the excellent obstacle-surmounting ability of traditional legged robots with the high-speed forward movement of wheeled robots, making them promising for applications such as field exploration, post-disaster search and rescue, and military reconnaissance. However, the contradictory issue of robot center of mass fluctuations and forward displacement during balanced C-leg movement poses a significant challenge. This paper proposes a reconfigurable Compliant wheel-leg based on environmental adaptability. Firstly, a modeling analysis is conducted, and the advantages of Compliant wheel-leg over traditional C-leg are verified through ADAMS dynamic simulation analysis. The dynamic simulation demonstrates that the Compliant wheel-leg can increase the robot's single-step forward displacement while ensuring that the robot’s center of mass fluctuation remains unchanged. This improvement not only ensures smooth and high-speed robot motion but also improves the robot's ability to cross trenches. Additionally, the Compliant wheel-leg can effectively reduce ground impact, enhancing the robot's environmental adaptability.

This paper presents an untethered pneumatic soft robot which can crawl both in horizontal and vertical pipes with different sizes and cross sections. This robot uses modular origami inspired soft-rigid hybrid actuator to produce telescoping and anchoring movements powered by vacuum pressure. The introduction of grooves to valley crease significantly lowers the full contraction vacuum pressure and improves the response, allowing the system can be driven by an onboard micro vacuum pump, enabling the possibility of miniaturization, integration, and untethered operation of the robot. A series of crawling experiments in pipes with different sizes and cross sections constructed by acrylic are conducted to validate the crawling performance of the robot. Within square cross-section pipes, the robot can achieve a velocity of 9.4 mm/s in horizontal crawling and 7.7 mm/s in vertical upward crawling. For horizontal crawling in circular pipes, it can reach a velocity of 8.0 mm/s. When fully charged, the robot can crawl for 40 min with a mileage of 16.649 m, which is sufficient for most drainage and industrial pipelines detection tasks. The robot demonstrates excellent endurance and speed performance that exceed most existing untethered soft pipe crawling robots.

Flapping wing propulsion is characterised by high efficiency and mobility, and many scholars have conducted in-depth studies on birds, manta rays, and other creatures with fluttering wing movements in anticipation of producing robots with high bionic effects. For wing-fluttering motion, the amplitude and frequency of fluttering are the key parameters affecting its effect. In this paper, a manta ray imitation robot is fabricated with manta rays as the research object, and the resonance of the ray bionic robot at the optimal flap frequency is realised by adjusting its intrinsic frequency with the flap frequency under the optimal working condition, so as to achieve the purpose of reducing the power and improving the propulsive efficiency. The two-dimensional two-degree-of-freedom bionic flap equations of motion were also established, and the flap simulation was carried out by using Ansys Fluent, and the relationship curves between the propulsive efficiency and Strahl's number (St) were plotted, and the optimal working conditions in the simulation were obtained by the optimal propulsive efficiency on the relationship curves. The results show that the optimum propulsive efficiency is 28

Underwater propulsion technology is a critical aspect of underwater equipment. The paradigm of underwater propulsion primarily based on propellers diminishes propulsion efficiency due to the lateral induced flow. Addressing this issue, our work proposes a reciprocating straight propulsion underwater propulsion strategy based on a passive folding mechanism. This propulsion form enhances effective thrust, and improves the motion efficiency of the vehicle via adopting the proposed straight propulsion mechanism which eliminates lateral induced water flow. To achieve vector propulsion ability, a 4-PRU parallel reconfigurable mechanism is added to the straight propulsion mechanism. The propulsion principle, parallel mechanism design, overall structure design of the vehicle, and dynamic analysis are detailed presented. Computational fluid dynamics analysis is used to validate the feasibility of the propulsion principle and the reliability of the dynamic model. Experiments confirm that the motion efficiency of the vehicle under uniform acceleration propulsion mode ranges from 61 % to 69 %, which is higher than traditional propeller propulsion methods. Additionally, the vector-propelled vehicle demonstrates good maneuverability in turning tests in a water pool.

The control of continuum robots is generally based on a kinematic model consisting of the actuation-configuration-task spaces. The nonlinear relationship between the actuation and configuration space makes it challenging to build control systems for continuum robots. This paper introduces the cyclotomic-linked kinematic model (CLKM), a pragmatic modeling approach designed to bridge this integration gap by replacing the configuration space with an intuitive joint space. The cornerstone of CLKM is the establishment of a linear relationship between the actuation space and the joint space, which crucially enables the direct application of standard robotic software stacks like ROS, MoveIt, and KDL for control and motion planning. In addition, the CLKM allows a uniform control architecture for continuum modules by using rigid joints, which makes it easy to model the kinematics of rigid-flexible hybrid arms. To validate this approach, a CLKM-based control system of a rigid-flexible hybrid arm prototype is constructed, which mainly utilizes software packages that have been widely used on discrete-jointed robots. The system seamlessly leverages off-the-shelf motion planners from the ROS ecosystem, and experimental results in path tracking tasks demonstrate its practical effectiveness, achieving a maximum tracking error below 1.5% of the continuum module's length.

Continuum robots, known for their compliance in unstructured environments, face limitations due to the lack of rotational degrees of freedom (DOFs) about the backbone. This prevents them from compensating undesired torsional deformation and performing 6-DOF control of the end-effector, thereby restricting their mobility. This paper presents a continuum robot with integrated dual rotational DOFs. One is integrated at the arm base to compensate for torsional deformation caused by external loads, while the other one, located at the arm tip, enables full 6-DOF control of the end-effector. To control the robot, a screw-theory-based kinematic model and a kinematic control framework are proposed to enable real-time, simultaneous control of the end-effector’s position and orientation. Experimental results show that the arm base rotational joint can fully compensate for undesired torsional deformation caused by a 1000 g payload. Thanks to the arm tip’s DOF and the proposed kinematic control framework, the robot’s end-effector can maintain a constant orientation while achieving open-loop path-tracking errors of only 3.3% of the arm’s length (930 mm), and successfully executing valve-closing tasks with coordinated 6-DOF motion, demonstrating the robot’s potential for industrial maintenance, human-robot interaction, and confined-space manipulation.

Grasp generalisation ability and intra-palmar manipulation ability are important functions of robotic hands. Currently, most robotic hands with these two abilities have a human-like structure with an excessive number of degrees of freedom, results in a complex and costly design that can hinder their practical application in scenarios prioritizing cost-effectiveness and simplicity. In response to this issue, this paper presents a new design for a robotic hand with three fingers and two degrees of freedom, capable of grasping and intra-palmar manipulation. This streamlined motion style, which combines a grasping degree of freedom and a twisting degree of freedom, can perform most hand grasping actions as well as intra-palmar manipulation without the need for wrist compensation. A theoretical model of the mechanism's motion and a force coupling model have been established, as well as a prototype. The robotic hand performance evaluation experiments and the force coupling compensation experiments are implemented. Experimental results verify that the robotic hand has strong grasping ability, fine intra-palmar manipulation ability and large load capacity. The proposed robotic hand provides a new solution for manipulation tools.

Capturing dynamic targets is particularly challenging for either rigid or soft grippers, as impact buffering should be completed in a short time to ensure the reliability of the robotic system. At collision onset, to deal with relatively low contact forces, adopting low stiffness and damping can effectively mitigate the rebound of the dynamic targets. As the contact area and forces increase, employing high stiffness and damping becomes necessary for absorbing high energy. This paper proposed a novel robotic gripper whose stiffness and damping follow a predefined profile "low stiffness and damping for low impact and high stiffness and damping for high impact." The variable effects of impact buffering and energy dissipation in a collision process were modeled and analyzed. Then, a passive variable stiffness and damping regulator (P-VSDR) was developed where tendons and pulleys are used to generate a nonlinear motion from a linear spring-damper unit. The contact dynamics model of the robotic gripper equipped with P-VSDR was established. Simulated and experimental results show that this gripper enables reliable capture of dynamic targets with different velocities.

Activation of the induced receptor for advanced glycation end products (RAGE) leads to initiation of NF- κ B and MAP kinase signaling pathways, resulting in propagation and perpetuation of inflammation. RAGE-knockout animals are less susceptible to acute inflammation and carcinogen-induced tumor development. We have reported that most forms of tumor cell death result in release of the RAGE ligand, high-mobility group protein 1 (HMGB1). We now report a novel role for RAGE in the tumor cell response to stress. Targeted knockdown of RAGE in the tumor cell, leads to increased apoptosis, diminished autophagy and decreased tumor cell survival . In contrast, overexpression of RAGE is associated with enhanced autophagy, diminished apoptosis and greater tumor cell viability. RAGE limits apoptosis through a p53-dependent mitochondrial pathway. Moreover, RAGE-sustained autophagy is associated with decreased phosphorylation of mammalian target of rapamycin (mTOR) and increased Beclin-1/VPS34 autophagosome formation. These findings show that the inflammatory receptor, RAGE, has a heretofore unrecognized role in the tumor cell response to stress. Furthermore, these studies establish a direct link between inflammatory mediators in the tumor microenvironment and resistance to programmed cell death. Our data suggest that targeted inhibition of RAGE or its ligands may serve as novel targets to enhance current cancer therapies.

Tensegrity structures have the advantage of superior deformation ability and high load-to-weight ratio, making them potential candidates for cable-driven continuum robot design. However, designing a clustered tensegrity continuum robot is still challenging due to the difficulty in modeling the tensegrity structure. In this study, we propose an innovation design method for a clustered tensegrity continuum robot based on machine learning (ML). Our ML-driven design method includes topology design of the clustered tensegrity continuum robot using genetic algorithm (GA) and deep reinforcement learning (DRL) approach, and driving law design (motion planning) of the continuum robot using deep reinforcement learning. We emphasize the obstacle avoidance and reaching point motion as one of the most important challenges for a cable-driven continuum robot and design the topology and driving law of the clustered tensegrity continuum robot through ML-based approach. This study demonstrates the applicability of tensegrity structures in the field of clustered tensegrity continuum robot design and illustrates the feasibility of using machine learning in the design of clustered tensegrity continuum robot.

Equipping advanced intelligent machines with human-like nervous systems requires new sensing materials capable of conformal integration into complex multidimensional structures. Here, we introduce an ionic liquid-enhanced polyacrylonitrile/AgNO3 formulation (IL-PANSion)—designed to be fabricated in diverse forms, including air-spun fibers, multi-material 2D printed layers, and 3D injection-molded networks. By leveraging room-temperature air spinning, IL-PANSion fibers achieve high tensile strain (up to 300 times that of conventional polyacrylonitrile fibers) and maintain over 650

The majority of underwater vehicles currently use screw propellers as propulsion method. Despite screw propellers are being promoted to increase the efficiency through better designs and technologies, the circumferential flow caused by the rotation of screw propeller results in wasted energy. This paper proposes an ideal reciprocating straight propulsion paradigm that eliminates this waste by implementing straight backward thrust based on a foldable mechanism. This mechanism can implement a maximized propulsion during the propelling stroke and a minimized resistance during recovery by folding and unfolding the mechanism, which can be adjusted passively via the relative motion between the thruster mechanism and water avoiding using sensors and actuations. This principle makes the reciprocating straight propulsion simple and reliable to be achieved in practice. Besides, a modularized propulsion system with four identical independently thrusters based on the proposed mechanism was proposed, achieving various motion modes in underwater vehicles. The proposed propulsion mechanism was theoretically analyzed and verified through simulations and prototype tests. In mooring tests, the thrust provided by the proposed mechanism is similar to the simulation results. The propulsion performance indicator of an underwater vehicle with the proposed propulsor reached 66 %. The mechanism could turn even with one-sided propulsion.

Most pneumatic actuators used in robotics are controlled by valves that contain moving parts (e.g., spool or rotor) and electronics to change the direction or pressure of the air flow. Thus, the dynamic bandwidth and robustness of the system are limited by these elements. This article presents an oscillation-based pneumatic actuation method to remove the moving parts and electronics from the valve. The obtained bistable load-switched (LoS) oscillator utilizes two output attachment walls to generate the Coanda effect and internal flow field to control the pressure in different output channels. The bistable LoS oscillator is implemented on a soft fish and runner, achieving locomotion speed up to 1.68 and 1.97 BL/s (body length per second), respectively, which are faster than existing counterparts. Furthermore, a single-output LoS oscillator is demonstrated by slightly modifying the bistable one. It enables the development of a soft runner with higher load capacity, as well as a relief valve used for pressure regulation in soft robotic grippers. The presented actuation methods can be potentially extended to a variety of situations that require compact size, light weight, high dynamics, and robustness.

Load-bearing walking is a demanding activity for individuals engaged in field exploration, rescue operations, and military tasks. This study introduces a non-anthropomorphic passive lower-limb exoskeleton (NAPLE) aimed at augmenting human load-bearing capacities. NAPLE adopts a reconfigurable universal-cylindrical-revolution pair (U-C-R) mechanism in each leg to passively accommodate the walking gait, thereby circumventing the misalignment limitations of motion between humans and robots widely existing in anthropomorphic exoskeletons. The topology of the mechanism can be reconfigured to a U-R mechanism via a cylindrical pair regressed in the stance phase and recruited in the swing phase, which is intelligently and automatically regulated via a purely mechanical approach. Meanwhile, a gravity compensation spring was proposed to smooth the fluctuation of the center of mass of the load to further decrease the consumption of the wearer. Prototype testing demonstrated that NAPLE can transfer an average of 87.8 % of the load weighing 30 kg to the ground in the stance phase and an average of 42.4 % while walking.

Load-bearing exoskeletons are expected to have high potential for use in military, fire-fighting, and logistics applications. Passive exoskeletons are more advantages owing to their light weight and long endurance. The load-bearing ability of such exoskeletons depends on the mechanical performance of joints, and reasonable output in knee joint is helpful to improve the load-bearing ability. In this article, to improve the load-bearing performance of the lower limb exoskeleton, a novel bianisotropic damper based on a magnetorheological fluid (BaDMr) is proposed, which is of bidirectional anisotropy in damping via a double-channel structure design to match the needed torque in exoskeleton knee during human walking with heavy load. It is of compact structure, light weight, and large torque output range. The ability of damping adjustment and load-bearing of the proposed system were verified via numerical simulation and a series of physical experiments. The results confirmed that BaDMr has a hysteresis rate of 6.42%, an average step response time of 16.06 ms, and torque tracking accuracy with a 5.3% root-mean-square error of the desire torque. In addition, an experimental platform of lower limb quasi-passive load-bearing exoskeleton was built to investigate the performance of BaDMr in real application scenarios. The load-bearing experiments verified that the exoskeleton could significantly reduce the impact of load on the wearer.

Dynamic modeling for continuum robots remains challenging due to their large nonlinear deformation and the variation of dynamic parameters during movement. In this paper, a lumped-mass dynamic model (LMD) for a continuum robot is constructed including elastic and viscous parameters in the robotic joints. Then the appropriate dynamic parameters (e.g. spring and damping coefficients of the LMD) with respect to the motion status (e.g. position and velocity of the robot) are estimated using a Genetic Algorithm (GA). Based on the obtained data set, a Multi-Layer Perception (MLP) is trained to establish a direct mapping from the motion status to the dynamic parameters, so the LMD can tune its parameters in real-time when moving within the workspace, resulting an adaptive lumped-mass dynamic model (ALMD). Compared to the fixed-parameter LMD, the modeling error of the ALMD is reduced by up to 60.2 %. Finally, a feedforward controller is implemented to control a continuum robotic prototype using the presented ALMD, reducing the maximum tracking error by 67.5 %.