Icing is a common natural phenomenon. However, it is widespread in various fields, including aerospace, power facilities, and others, where it can lead to equipment failure, reduced energy efficiency, and even severe safety accidents. In recent years, biomimetic superhydrophobic materials inspired by the surface of lotus leaves have shown significant potential for applications in anti-icing and de-icing due to their excellent hydrophobicity and low surface energy. This paper systematically introduces the anti-icing mechanisms of superhydrophobic coatings and provides a review of recent research progress on superhydrophobic materials in the field of anti-icing. Additionally, the application of these materials in three scenarios: wind turbine blades, power transmission lines, and aerospace equipment, was analysed to assess their prospects and feasibility in engineering projects. Finally, the current status of superhydrophobic anti-icing materials and the challenges they face are summarised, highlighting existing issues and offering perspectives for future research directions.

This review explores the potential of incorporating Machine Learning (ML) and Artificial Intelligence (AI) into the control systems of Lower Limb Rehabilitation Exoskeletons (LLREs), with a focus on its capacity to advance the field of rehabilitation robotics. It examines applications in optimizing personalized control supporting trajectory adaptation and key metrics in rehabilitation and discusses the challenges and limitations of traditional ML control methods supervised, unsupervised, and Reinforcement Learning (RL) in LLRE control. We analyze the specific aspects of traditional ML techniques in controlling the rehabilitation exoskeleton. Finally, the results of scientific solutions and developments in the implementation of gait parameters for personalized control that support the trajectory adaptation of the rehabilitation process are studied. Clinical applications and case studies directly describe the shortcomings and advantages of the field.

Biological systems such as mountain goats and felines exhibit remarkable agility and adaptability when traversing complex terrains. Inspired by these capabilities, quadruped robots have been developed to mimic legged locomotion and improve mobility over uneven environments. To further enhance locomotion efficiency and terrain versatility, wheeled-legged robots integrate wheels and legs into a hybrid platform, enabling both high-speed traversal and robust ground contact in unstructured terrain. However, planning coordinated locomotion across diverse terrains remains challenging due to the nonlinear dynamics, complex terrain contact constraints, and multimodal locomotion capabilities. In this paper, we propose a real-time, integrated planning framework that jointly optimizes gait scheduling, footstep placement, and whole-body motion trajectories. Our method adopts a two-stage approach. First, a sampling-based planner generates candidate gait sequences and nominal footstep targets based on terrain features and kinematic feasibility. Second, a constrained trajectory optimizer reformulates the planning problem as a Quadratic Programming (QP) task to compute dynamically feasible base trajectories and corresponding ground reaction forces. This hybrid formulation balances planning efficiency and physical realism. The planned trajectories and contact forces are tracked using a hierarchical control architecture combining Model Predictive Control (MPC) and Whole-Body Control (WBC), enabling fast and stable execution on real hardware. Simulation and real-world experiments demonstrate that our approach enables adaptive gait transitions and improves terrain adaptability compared to traditional planners.

The load capacity and the environmental adaptation of insect-scale robots are important for their practical application. In this study, a piezoelectric driven Parallel-Legged Insect-scale Modular Robot (PLimBot) is developed. The bot features sub-gram weight (0.96 g) and insect-scale dimensions (3.5(L)?×?2.5(W)?×?2.7(H) cm) with a high chassis configuration. PLimBot can achieve a maximum velocity of 3.3 BL/s without load, and 0.3 BL/s with a 10 g on-board load (>?10 times body weight). By leveraging its geometric configuration, the bot can adapt to multi-type obstacles in constrained environment, pass sharp turns with loads (1 g), and navigate through intricate mazes. PLimBot can climb a 20° slope at 0.57 BL/s, navigate through gravel surfaces, multi-medias, and shoals measuring up to 2 cm in height (0.74 BH). We anticipate the PLimBot can provide a new method for the design, fabrication, analyse, and benchmarking of insect-scale micro-robots.

Hexapod robots demonstrate significant advantages in payload capacity and stability, making them an excellent choice for complex terrain applications such as disaster rescue and mountain transportation. However, operating on low-friction surfaces often leads to foot slippage, risking instability and performance degradation. In this work, we propose a novel integrated slip detection and recovery framework for actively force-controlled hexapod robots. We first analyze the fundamental characteristics of foot-ground interaction, revealing that slippage arises when the desired foot force exceeds the actual friction limits due to inaccuracies in estimated ground parameters. Based on this analysis, we introduce a velocity-based slip detection strategy, enabling prompt identification of slip events. Upon detection, we employ a Back Propagation (BP) neural network to accurately re-identify ground parameters, including unit normal vectors and friction coefficients at the contact points. These refined parameters are then integrated into a Quadratic Programming (QP) framework via an interpolation scheme, ensuring smooth transitions in foot force distribution. Both simulation and physical experimental validation demonstrate that our framework successfully detects slippage events, accurately identifies ground parameters, and achieves foot stabilization through corrected force distribution. This work offers a comprehensive solution to the mismatch between estimated and actual ground parameters, enabling hexapod robots to achieve stable locomotion in challenging low-friction environments.

Pulsed jet propulsion, a highly efficient locomotion strategy commonly observed in marine organisms, offers significant potential for bioinspired underwater robotics. Drawing inspiration from the jet propulsion mechanism of squids, this study presents the design and implementation of a soft biomimetic robotic fish powered by liquid metal-based electromagnetic actuators. The robot employs a novel actuation strategy in which Lorentz forces are generated via the interaction between embedded permanent magnets and energized liquid metal coils, inducing periodic contraction and relaxation of a flexible body cavity to achieve efficient pulsed jet propulsion. Through comprehensive analysis of actuator dynamics and systematic optimization of structural and control parameters, the robotic fish attains a maximum linear swimming speed of 1.71 BL/s (Body Lengths per second), equivalent to 8.9 cm/s. To enhance maneuverability, a dual-cavity independently controlled architecture is employed alongside an inclined nozzle design, enabling precise directional control with a maximum turning rate of 47.4 deg/s and a minimum turning radius of 0.4 BL. Furthermore, the successful integration of miniaturized power and control systems within the body enables fully untethered autonomous swimming, greatly enhancing the system’s practicality. Experimental results demonstrate that the proposed robotic fish combines a compact structure with rapid response, versatile control, and excellent locomotion performance, highlighting its strong potential for applications in marine exploration, underwater surveillance, and environmental monitoring.

In this paper, a novel wire-driven robotic fish is proposed and developed. The innovation lies in achieving reciprocating tail fin oscillation through continuous motor rotation, with adjustable oscillation amplitude. The robotic fish is capable of swimming without frequent motor reversals and can dynamically adjust its tail fin oscillation amplitude and frequency to adapt to different environmental demands. First, the paper presents a detailed description of the robotic fish’s mechatronic design and working principle. Second, a comprehensive kinematic analysis is conducted. The dynamic model of the robotic fish is established using the Newton-Euler method, and the hydrodynamic parameters in the model are optimally identified through a data-driven parameter identification approach. To validate the performance of the robotic fish and the accuracy of the dynamic model, relevant experiments and simulations are conducted, and their results are compared. The simulation results demonstrate that the robotic fish exhibits excellent swimming performance and reveal the effectiveness of the dynamic model. Finally, field experiments are conducted, demonstrating the robotic fish’s good swimming stability and environmental adaptability. The research in this paper provides insights for enhancing the practical value of robotic fish in real-world applications.

Many aquatic species exhibit the extraordinary ability to leap out of water, with flying-fish being a prime example, capable of gliding impressive distances through the air after breaching the surface. Inspired by this remarkable mode of locomotion, researchers have pursued innovative designs for amphibious robots capable of seamlessly traversing the water-air interface. In this study, we present the development of an untethered bioinspired flying-fish robot actuated by the reciprocating motion of its caudal fin. The robot features retractable and rotatable pectoral fins on both sides, allowing active pitch control during submerged swimming. To enhance propulsion performance, we developed a dynamic model that describes the coupled head-tail beating motion and conducted a numerical optimization of the beating angles using FLUENT simulation. The results indicate that increasing the tail beating angle substantially boosts thrust albeit at the cost of greater lateral force, whereas increasing the head beating angle slightly reduces thrust while exerting minimal influence on lateral force. An optimal configuration, consisting of a 15 head beating angle and a 35 tail beating angle, achieves a balance between high thrust and low lateral force. Under these parameters, the robot attains a simulated swimming speed of 1.3 m/s and a jumping height of 86.14 mm, sufficient to breach the water surface. Experimental validations demonstrated a maximum swimming speed of 1.12 m/s, a jumping height of 112 mm, and a peak thrust of 9.819 N at a maximum beating frequency of 12 Hz. The robot also exhibits high maneuverability, with a minimum turning radius of 35 mm. Comparative analysis of simulation and experimental results shows relative errors of 16.07% in speed, 23.11% in jumping height, and 4.51% in thrust, with overall trends consistent between the two. These findings offer valuable insights for the design of next-generation amphibious robots capable of efficient water-to-air locomotion.

Due to their lightweight and flexibility, soft biomimetic robots are popular in deep-sea exploration. However, existing buoyancy materials lack optimal compatibility. This study proposes a flexible, pressure-resistant, multi-medium buoyancy module comprising a flexible cavity filled with a Hollow Glass Microsphere (HGM)-water mixture and introduces structured-grid thinking, which enables contour adaptation to complex biomimetic robot morphologies. The density and pressure resistance of the buoyancy modules were experimentally tested, and the effects of varying silicone hardness, wall thickness, and volume percentage of HGM in the mixture on the performance of the buoyancy modules were compared. The results indicate that the density of the buoyancy modules ranges from 0.751 to 0.964 g/cm3. Under a pressure of 30 MPa, the volume change rate of the buoyancy modules is between 1.74% and 2.13%. The effect of air content in the flexible cavity on buoyancy modules under high pressure was examined by comparing experimental findings with simulations.

Current directional control methods for flexible piezoelectric film robots often introduce additional structural complexity. Inspired by the lateral undulation of snakes, this paper presents a frequency-based directional control method for a robot with a monolithic structure. This research introduces a soft robot based on a helical composite piezoelectric film, innovatively designed to utilize a single spatial helical structure to achieve multidirectional foot motion. This design enables control over the direction and mode of motion via frequency modulation. A change in driving angle from ??19° to 123° is attained through frequency modulation, while a maximum linear motion speed of 3.4 BL/s (Body Length/s) is achieved. Empirical findings robustly demonstrate the robot’s ability to execute various tasks, including slope ascent, load carriage, maze traversal, and pipe navigation, thus showcasing its exceptional operational capacity within confined environments. Diverging from conventional paradigms that seek to enhance control through structural complexity, this groundbreaking study introduces a frequency-driven control methodology for flexible robots. This paradigm shift holds significant implications for streamlining structural designs, optimizing control strategies, and expanding the application scope of soft robotic systems.

Biologists have calculated the allometric growth patterns of birds in terms of morphological and motion parameters, expressing the five parameters of wingspan, wing area, cruising speed, flapping frequency, and flight power as power functions of weight. Aviation designers utilize this formula to derive the initial values of the morphological and motion parameters of Flapping Wing Air Vehicles (FWAV), a process we refer to as the traditional scaling method. Traditional avian scaling laws establish power-exponential relationships between body weight and key flight parameters, guiding bionic aircraft design. However, limitations persist in their direct application to FWAVs. This study proposes an enhanced scaling law methodology addressing four critical aspects: (1) Introducing a weight correction coefficient to map avian flight functionality specifically to FWAVs design requirements, isolating flight-related mass from total avian mass associated with broader biological activities. (2) Rectifying the frequency formula by incorporating wing length-frequency statistics and integrating flapping amplitude via the Strouhal number definition, comprehensively representing motion parameters through periodic average angular velocity. (3) Refitting scaling law parameters using morphological data from specific bionic objects (e.g., pigeons, eagles) to reduce parameter scatter inherent in traditional laws derived from birds spanning grams to kilograms with divergent morphologies. (4) Validating the method through the design and testing of a pigeon-inspired FWAV prototype. Experimental results demonstrate that parameters estimated via the improved method align closely with the optimized aircraft’s performance, achieving an endurance of 185 min on a single charge. This approach significantly reduces intermediate optimization needs, enhances design rationality, and provides a novel pathway for efficient FWAV development.

Despite advancements, amphibious robots still exhibit inadequate locomotion performance. As natatores demonstrate exceptional walking and swimming capabilities, they provide an ideal bionic prototype for structural enhancement. Consequently, this study designed a novel propulsion mechanism inspired by waterfowl, featuring a slider-four-bar linkage system that simultaneously enables foot folding and retracting. Static structural analysis was performed to identify and optimize weak components of the mechanism. The achievable workspace of the mechanism was calculated through kinematic analysis. Mathematical and numerical hydrodynamic models were established to evaluate the mechanism’s hydrodynamic performance using a defined thrust efficiency metric. To analyze the influence of kinematic parameters on the propulsion mechanism across full motion cycles, a surrogate model correlating kinematic parameters with the thrust efficiency metric was developed. Results demonstrate that the optimized propulsion mechanism satisfies the requirement of static performance under working conditions. Compared to the Waterfowl-inspired robot-I (WIR-I), the proposed mechanism achieves reduction in retraction stroke resistance for robotic webbed-foot propulsion systems. Surrogate model results show the following hierarchy of influence across kinematic parameters on the thrust efficiency metric: Paddling angle per cycle exceeds temporal baseline, which in turn exceeds temporal scaling factor of extension phase, power stroke, flexion phase, and retraction stroke. Flow field characterization indicates that flow attachment constitutes the dominant factor governing thrust efficiency. These findings could provide new design principles for future biomimetic webbed-foot robotic systems and establish a theoretical foundation for hydrodynamic optimization in amphibious robotics development and deployment.

Mirror-assisted strategies are commonly used in the rehabilitation training of patients with hemiparesis in the upper limbs following a stroke. Traditional robotic mirror assistance focuses on achieving high-precision mirror trajectory tracking, often neglecting the issue of active movement in the affected side. This paper proposes a task performance-based adaptive impedance control, where the robot assists the affected side in an assist-as-needed manner, thereby encouraging the patient to perform active movements. To account for inter-individual variability, a method for assessing the affected side’s motor performance, based on the healthy side’s movement level, is introduced. Adaptive impedance control is then constructed based on the motor performance of the affected side, enabling the robot to provide adaptive assistance force. Eight healthy participants were recruited for experimental testing. Experimental results show that when the robot provides mirror-based assist-as-needed to the affected side, the robot’s stiffness coefficient and assistance force are positively correlated with the motor assessment coefficient of the affected side, thereby verifying the feasibility of the proposed strategy. This study offers a robotic-assisted rehabilitation strategy for stroke patients that balances active participation and individual adaptability, with the potential to enhance rehabilitation outcomes and enable precise rehabilitation interventions.

Achieving high drive efficiency remains a significant challenge in active knee prosthesis design. Inspired by human knee biomechanics, this study presents a novel biomimetic hydraulic drive system integrated with three human-like mechanisms: antagonistic muscle driving mechanism, dynamic simulation of muscle forces, and multi-stage collaborative energy supply. The system features a multi-stage hydraulic-rope hybrid transmission enabling adjustable damping and compliant motion control, coupled with a dual-cylinder configuration that boosts driving efficiency while delivering 29.7 Nm peak torque. A pump-valve hybrid control strategy is developed to dynamically adjust the flow and driving torque across gait phases, enhancing response speed and angular tracking accuracy. Through computational modeling, simulation, and prototype validation, we demonstrate that the proposed hydraulic drive system achieves efficient and responsive knee flexion and extension while meeting functional demands, reducing energy consumption by 20–50% compared to traditional pump-controlled systems. This study introduces a novel strategy for developing multimodal muscle-joint collaborative mechanisms, establishing a foundational framework for next-generation, high-performance bioinspired prostheses.

Hydraulic artificial muscles offer higher stiffness and faster response. Unlike pneumatic artificial muscles, which rely on pressure control, hydraulic artificial muscles are more suitable for volume control. This study focuses on the modeling and nonlinear dynamic analysis of volume-controlled McKibben hydraulic artificial muscles. To address the limitations of existing models, a hierarchical semi-empirical modeling method is adopted to construct a static model with constant volume, which integrates a geometric model and a static force model. This model supports not only pressure control, but also volume control, and considers the end effects, the effective bulk modulus, and the nonlinear dependence of artificial muscles on Young’s modulus. Furthermore, a dynamic model encompassing four state parameters is established by combining the static model with fluid continuity equations, enabling the characterization of dynamic responses under volume control. Experimental methods were employed to determine the parameters in the formulas, significantly improving the accuracy of calculating the required injected volume for the given contraction ratio and load force. Nonlinear dynamics analysis reveals that HAM exhibit significant nonlinear damping, and the initial volume of connecting pipes significantly affects static pressure and working performance. Experimental validation shows that the static model achieves a maximum relative error of 2.05% in contraction ratio, while the dynamic model accurately captures the second-order oscillatory characteristics of force and pressure responses. This research provides a comprehensive modeling framework for volume-controlled HAM and deepens the understanding of their nonlinear dynamic behavior, facilitating engineering applications in robotics and rehabilitation devices.

Bio-inspired and bio-hybrid principles are used to design robots that imitate biological systems. Animals and insects are commonly selected as the prototype in this context due to their mobility. However, plants have been found to grow slowly but move rapidly based on hydraulics and mechanics. The Venus flytrap (Dionaea muscipula) is an example of a carnivorous plant that eats insects, but the mechanism underlying its quick predation behavior has not been adequately explored in research. It is particularly important to determine how cells of this plant are grouped together, from the level of the cell to that of the tissue as well as from the microscopic to the macroscopic scale, such that they can perform this motion without the help of muscles. In this study, we map the cellular characteristics and cell distribution of the Venus flytrap. We propose an area-weighted central point to describe a units-driven principle of actuation and develop a computational model to verify it. We fabricate and test a physical model that patterned the cellular structure of the Venus flytrap to explore this mechanism by using chemically fixed and immobile samples. The hydrogel-based actuator replicates the deformation of the plant cell under turgor pressure. The statistical results show that the area-weighted center of the plant involves a shift, and this distribution actuates the bending-induced deformation of its lobes.

Inspired by the loofah sponge’s axially continuous porous core and peripheral hexagonal scaffold, we propose a compact bio-inspired low-frequency vibration isolator. An analytical static model reveals a Quasi-Zero-Stiffness (QZS) region produced by parallel coupling of an axial positive-stiffness spring and a nonlinear hexagonal link–spring unit. A Lagrangian dynamic model and ADAMS multibody simulations predict resonance and transmissibility and are validated by sinusoidal base-displacement tests. With a 5.775 kg payload, the prototype achieves effective isolation above ~?5 Hz and ~?25–30 dB attenuation at 20–22 Hz while retaining comparable load capacity to linear references of similar size. Compared with linear isolators, the designed bio-inspired low-frequency vibration isolator exhibits a lower isolation onset and a broader useful bandwidth under a compact footprint, offering tunable low-frequency isolation via geometric and stiffness parameters (α, k₁, k₃, l).

Insects are exceptionally robust walkers. Although different species exhibit distinct anatomical and functional specializations, they are also highly adaptive within these constraints. How such adaptations enable insects to efficiently navigate diverse environments and perform mechanical tasks remains far from fully explored. The mole cricket, which dwells underground, has been little studied to date, largely due to its cryptic lifestyle. It excels at digging tunnels and exhibits extreme morphological adaptations, particularly its exceptional fossorial forelegs. Its versatile locomotion, above and below ground, makes the mole cricket an attractive model system for studying the biomechanics of insect movement. Here we provide the first quantitative characterization of mole cricket locomotion. Using a tunnel-like arena, we video monitored insects and analyzed their locomotion gaits. We identified and described three main modes of locomotion, including a new, recently described backward-bound gait. To test specific hypotheses regarding form-function relationships and the generation of thrust, we integrated biomechanical modeling and deep reinforcement learning to simulate the observed gaits. Our work opens several future directions, from exploring context-dependent gait transitions to bio-inspired technological innovations.

Gesture classification based on surface electromyography (sEMG) have been extensively investigated towards the control of smart prosthetic hands. However, most studies did not consider the effects of arm position and movements that frequently occur in daily activities. This study aims to address the gesture classification challenge under arm movements. We collected sEMG and acceleration (ACC) data from fourteen participants, including two individuals with radial artery amputations, while performing gestures in both static and dynamic arm states. Using the collected data, the performance of three machine learning methods was evaluated for gesture classification under both arm movements and static arm conditions. The results revealed a 17.48% decrease in average classification accuracy in the dynamic state compared to the static state when using sEMG signals. Subsequently, the improvement in classification accuracy under arm movements was validated using both sEMG and ACC, with deep learning achieving the highest average accuracy of 84.35% across healthy subjects. Additionally, the study assessed the impact of gesture similarity on classification performance and evaluated the practical efficacy of classifiers for amputees. To further enhance gesture classification accuracy under arm movements, a two-stage gesture classification model training method based on ResNet18 was proposed. This method first learns a generalized motion prototype from population data and then adapts it to individual subjects via fine-tuning, resulting in a 2.49% improvement in average recognition accuracy across all 14 subjects.

The pain induced by the friction between residual limb and prosthetic socket is the significant issue that influence the comfort of prostheses. In this study, to reveal and compare the pain generation and processing mechanisms in amputees and healthy individuals, the skin friction, electroencephalogram, and functional Near-Infrared Spectroscopy tests were carried out. The Hodgkin-Huxley and Gate Control Theory models were used to reveal the neural transduction and transmission of frictional pain and connect the surface friction and brain activation of pain. The skin friction and subjective evaluation tests showed that compared with healthy skin, the residual limb skin exhibited poorer elastic properties, lower friction coefficients and higher frictional pain thresholds. The results showed that the hemodynamic response of amputees was insignificant and brain regions of amputees related with γ oscillations was smaller compared with healthy individuals. The changes of skin mechanical properties and neural reorganization may responsible for the difference in frictional pain response between amputees and healthy individuals. The results also showed that the increasing intensity of frictional pain can induce the increase of T-cell potential, resulting in the increasing HbO concentration and γ oscillations. This study provides a better understanding of pain generation mechanisms from skin surface to brain response using mathematical models. It also reveals the differences in frictional pain between amputees and healthy individuals.

Inspired by bionic surface design principles, this study developed a one-step method to deposit thiol-isocyanate polymer microspheres on dopamine-treated carbon fiber surfaces, creating a raspberry-like surface architecture. The bioinspired structure serves as both physical and chemical “bridges” between carbon fibers and epoxy resin, while simultaneously enhancing the chemical activity, mechanical interlocking effect, and wettability of the fiber surfaces. Compared to untreated samples, the modified composites exhibited 63.11% improvement in interlaminar fracture toughness, with impact strength and tensile strength increasing remarkably by 167.55% and 73.59%, respectively. Furthermore, systematic investigation of the interfacial reinforcement mechanisms and energy dissipation modes was conducted through comprehensive fracture morphology analysis.

Cutting tools for extracting natural latex from wood have been used for centuries, yet research on tapping knives remains scarce, particularly regarding efficient surface wood cutting tools. This study designed two rotary blades biomimicking woodpecker beak angles for wood-cutting tests. Tests examined surface quality and blade wear under varying parameters. Results demonstrated that the Beak-Edge-Blade was optimal for low-speed (5 mm/s) parallel-to-grain cutting based on peak cutting forces and energy consumption, while the Tip-Beak-Blade performed better in high-speed (15 mm/s) parallel and cross-grain cutting. Both blades showed comparable performance at medium speed (10 mm/s). TiCN coating effectively reduced blade wear but increased cutting forces and energy consumption at both low and high speeds, while reducing these parameters at medium speed. Parallel-to-grain cutting achieved better surface integrity (Sa, Sz, Str) than cross-grain. EDS/SEM revealed two wear mechanisms: fiber adhesion and edge chipping. Beak-Edge-Blade suffered chipping; Tip-Beak-Blade exhibited adhesion. Findings aid precision tool design, notably depth-sensitive tools like rubber-tapping blades.

The rising prevalence of hip joint disorders, particularly among aging populations, highlights the need for advanced surgical and rehabilitation strategies. The artificial ligament plays a key role in restoring joint stability in the treatment of hip joint disorders. Existing commercial artificial ligaments differ from biological ligaments in that they lack the complex hierarchical organization of natural ligaments. This study introduces bioinspired hierarchical 3D braided ligaments to replicate the nonlinear mechanical behavior of human ligaments. We examined the effects of braiding strands and angles on the tensile properties of artificial ligaments, including toe-region strain and linear modulus. Key characteristics such as stress relaxation and fatigue were also assessed. Using FEA, we simulated fiber interactions and macroscopic mechanical behavior, revealing the mechanisms behind the J-shaped curve of braided ligaments. Based on theoretical analysis, we selected a high-fidelity artificial braided ligament and compared the hip joint’s range of motion with and without it. The results show that the artificial hip with the round ligament closely mimics the human hip’s motion (beyond 95% similarity in all directions including three translations and three rotations), which reveals their potential to enhance joint stability and serve as effective therapeutic and educational tools in medical practice.

Soft robots, owing to their ability to undergo continuous large deformations, hold significant promise in fields such as disaster rescue, scientific exploration, and medical healthcare. However, they are prone to damage under harsh working conditions, which can negatively impact their operational efficiency. The existing research on the protection of soft robots is very limited. In this study, inspired by the natural armor of the arapaima and the asymmetrical body surface structures of the cobra, a high-performance flexible protective system was designed using a coupled biomimetic approach. An electronic pressure testing machine was used to evaluate the flexibility and protective performance of the biomimetic protection system. The results indicated that the presence of the asymmetric structures can substantially enhance the protection performance of the biomimetic protection system without appreciably compromising its flexibility in the slightest. This is due to the reason that the frictional anisotropic effect of asymmetric structure has varying effects on the interlocking interaction between adjacent scales in bending and compression conditions. This study introduces a novel approach aiming at enhancing the protective performance of flexible protection systems, thus presenting potential applications in soft robotic systems.

Osteomyelitis caused by Staphylococcus aureus (S. aureus) is a severe inflammatory bone disease that is difficult to eradicate and can be life-threatening. Traditional treatments relying on high-dose systemic antibiotics often fail due to biofilm resistance and emerging drug-resistant strains. This study proposes a diflunisal-loaded 3D-printed scaffold as a novel therapeutic strategy. A composite biomaterial ink composed of gelatin (Gel), polycaprolactone (PCL), and nanohydroxyapatite (n-HA) was synthesized and used to fabricate customized porous scaffolds via 3D printing. Diflunisal was loaded onto the scaffolds using a Gel swelling method. In vitro experiments showed sustained diflunisal release under different pH conditions (pH?=?6.0 and 7.4) mimicking infection and protection of bone marrow stromal cells (BMSCs) from S. aureus toxins. In vivo studies revealed significant alleviation of infection and promotion of bone regeneration at the defect site. This diflunisal-loaded Gel/PCL/n-HA scaffold integrates anti-virulence therapy with bone regeneration, offering a promising solution for osteomyelitis treatment.

Monitoring concrete cracks for structural health in civil engineering presents a significant challenge. This is primarily due to the reliance on manual investigation methods, impacts of global climatic shifts stress, and geohazard threats to engineering structures. To cope with this challenge, state-of-the-art Deep Learning (DL) models are utilized to predict concrete cracks and accurately identify subtle variations in crack patterns and sizes, which lighting conditions and surface textures can influence. Previous studies indicate that model accuracy may decrease when faced with obscured concrete cracks, irregular shapes, or limited datasets for real-world problem scenarios. Feature fusion enhances model performance by combining complementary information, resulting in more accurate predictions, but may increase complexity and potential information redundancy. The study presents the Fractur Encoder to Decoder (FractED) block, a novel architecture consisting of three sub-blocks: the inner block (Encoder), intermediate block (Intermediate block), and outer block (Decoder). This approach integrates fused features into the model without additional fine-tuning steps, allowing for comprehensive feature refinement and enhancement, ultimately optimizing model performance. The study investigates a DL methodology on three datasets, demonstrating its effectiveness in handling complex classification scenarios in civil engineering. The model achieved high accuracy rates, with 88.41% for multiclass (Deck, Pavement, and Walls) classification tasks, 91.94% on the Pillow Dam Borehole image binary dataset, and 99.77% on the Surface Crack binary dataset. The FractED block integration ensures adaptability and scalability, making it valuable for various Artificial Intelligence (AI) applications in civil engineering. The research also provides a scientific foundation for automatizing civil engineering inspection instruments for the future.

Metaheuristic algorithms have emerged as indispensable tools for solving NP-hard optimization problems that defy traditional methods. To advance the field’s focus on algorithmic performance, this study introduces the Theory Evolution Optimization (TEO) – an efficient metaheuristic inspired by the evolution of scientific theory. TEO simulates the competitive, accumulative, and replacement processes among scientific hypotheses, mirroring the evolution from a hypothesis to an established scientific theory. The performance of TEO is validated through extensive experimental simulations and benchmarked against 28 popular algorithms, including highly competitive champions such as EBOwithCMAR, LSHADE_cnEpSi, and LSHADE. Pairwise comparisons between TEO and the latest algorithms are conducted using the Wilcoxon signed-rank test, with multiple comparisons managed by the Friedman test. Initially, TEO is tested on the classical IEEE CEC2017 and the latest IEEE CEC2022 benchmark functions. TEO successfully addresses four prominent engineering design problems in constrained continuous space for practical applications. Additionally, a binary TEO (BTEO) variant is introduced and applied to feature selection tasks in discrete space. Experimental results consistently demonstrate that TEO proposes highly competitive outcomes in optimization problems. The source codes for this research are accessible to the public at https://aliasgharheidari.com/TEO.html.

In this study, we present a novel approach to multi-threshold image segmentation using an adaptive method that combines the Ebola Optimization Search Algorithm (EOSA) with the Aquila Optimizer, termed the Integrated Enhanced Ebola Optimization Search Algorithm (IEOSA). Our approach leverages this integration to produce high-quality segmented images. The IEOSA method introduces two distinct optimization mechanisms to identify optimal solutions. By blending the randomness of the Aquila Optimizer with the capabilities of EOSA, we enhance the exploration potential of the algorithm. Additionally, we incorporate a self-transition learning system within the IEOSA to further boost its performance. To tackle multi-level threshold image segmentation, we apply Kapur’s entropy between-class variance within the IEOSA framework. Our findings show that the IEOSA-based techniques outperform other comparable methods, offering faster convergence and more stable segmentation results. Through comparative analysis using standard test images, we demonstrate that IEOSA achieves higher solution accuracy than other methods. Ultimately, the proposed IEOSA methodologies effectively address multi-level threshold image segmentation challenges, accurately segmenting even the minor errors that are often overlooked in high-resolution images.

The Marine Predators Algorithm (MPA), while promising for complex optimization, suffers from limited solution precision, imbalanced exploration–exploitation, and premature convergence. To address these shortcomings, this paper proposes a phased- enhancement variant named PEMPA, which integrates three novel strategies into distinct phases of MPA: 1) embedding historical best positions in the high-velocity ratio phase to refine solution quality; 2) introducing an adaptive inertia weight based on an inverted Sigmoid function in the unit-velocity ratio phase to systematically balance exploration and exploitation; and 3) designing a two-stage opposition-based learning operator in the low-velocity ratio phase to prevent premature convergence. The performance of PEMPA is comprehensively evaluated across 23 classical benchmark functions, the IEEE Congress on Evolutionary Computation (CEC) 2017 test suite, 21 feature selection tasks, and a real-world medical insurance fraud detection problem. Experimental results confirm that the proposed strategies significantly enhance the efficiency and robustness of MPA. Furthermore, PEMPA demonstrates highly competitive performance compared with several state-of-the-art metaheuristic algorithms, validating its effectiveness and scalability for diverse optimization challenges.

Researchers continuously advance optimization algorithms, recognizing that no meta-heuristic can solve all problem types, as stated by the No Free Lunch theorem. This paper introduces the Propagation alongside Crossover (PAC) algorithm to address continuous optimization challenges. The primary goal of PAC is to structure the algorithmic phases in a manner that achieves a robust balance between exploration and exploitation through appropriately designed mechanisms at each stage. PAC simultaneously leverages the benefits of propagation, crossover, and mutation. Three independent operators are defined to generate new candidate solutions separately, and a novel selection strategy allows individuals produced by each operator, along with members of the current population, to independently enter the next generation. This design preserves population diversity, prevents all individuals from converging toward a single point, and enhances the algorithm’s ability to explore the solution space effectively. A key innovation of PAC is its three-mode propagation mechanism, which comprises local search, linear propagation toward the target point, and tear-drop shaped propagation toward the target point. Tear-drop propagation provides a precise and adaptive search around promising solutions, increasing diversity and preventing entrapment in local optima. The target point is typically set as the global optimum; however, when propagating the global optimum itself, a random point is used as the target to further enhance exploration and escape from local optima. The initial population is generated using chaotic mapping to ensure broad coverage of the search space. PAC was rigorously evaluated on 51 benchmark functions and three engineering problems, considering scalability, convergence, sensitivity, and computational efficiency. Comparative analyses with established optimization algorithms demonstrate PAC’s superior performance, as confirmed by Wilcoxon signed-rank and Friedman statistical tests. Furthermore, PAC was applied as a feature selection method on four diverse datasets, achieving substantial dimensionality reduction while outperforming comparative methods in classification accuracy. These results highlight PAC’s versatility, robustness, and practical effectiveness.

Liver Fibrosis (LF) is a progressive liver disease that requires accurate staging for effective treatment. The conventional manual diagnostic methods are more time-consuming, labor-intensive, and susceptible to human error. Additionally, medical image datasets often contain noise and poor quality, which degrade Deep Learning (DL) model performance. Meanwhile, conventional aggregation operators often fail to fully leverage the unique strengths of individual DL models by highlighting the need for more robust aggregation methods. To address these challenges, we propose a novel ensemble network based on Dempster-Shafer theory-driven fusion (DSFuse-Net). The proposed approach begins with an advanced preprocessing phase where liver ultrasound images undergo contrast enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE) and noise suppression via Non-Local Means filtering. These enhanced images are then fused using Daubechies wavelet transform by leveraging both low- frequency and high-frequency decompositions to improve image quality and highlight critical fibrosis features. For robust feature extraction, we carefully selected three pretrained DL architectures including DenseNet201, DenseNet169, and ResNet50V2 for considering them as base models. These models serve as candidates for fusion using Dempster-Shafer theory (DST), which enables dynamically weights predictions based on confidence. The proposed approach effectively integrates evidence by handling uncertainty and conflicting information, thereby improving the overall decision-making process. We validated DSFuse-Net on a publicly available LF dataset and achieved superior accuracy of 98.42% among five LF stages labelled as F0 to F4. We evaluate model performance using key metrics and analyze results with confusion-matrices, Receiver-Operating-Characteristic (ROC) curves, and Precision-Recall (PR) curves to assess classification accuracy. Additionally, Grad-CAMs and SHAP (SHapley Additive exPlanations) are employed to provide visual explanations and interpret feature contributions. Furthermore, statistical significance is validated using McNemar’s and the Wilcoxon Signed-Rank test to confirm that the findings are not due to random chance.

Autistic Spectrum Disorder (ASD) is a neurodevelopmental disorder that leads to complexities in social interactions. Based on the statistics of the WHO, more patients diagnosed with ASD is steadily increasing. Most of the recent investigations focused on medical treatment and data collection, but they did not focus on ASD diagnosis related to Deep Learning (DL). To solve this issue, this work introduces an efficient model for ASD detection with multi-modal data utilizing Convolutional LeNet Forward Taylor network (CLeFTNet). Here, the CLeFTNet is developed by combining CNN and LeNet using the Taylor series. This work involves two kinds of inputs: First, an input brain image is pre-processed by an Adaptive Wiener filter (AWF) and Region of Interest (RoI) extraction. Then, the Box neighborhood search algorithm is employed for extracting the pivotal region regarding functional connectivity, and feature extraction is performed. Simultaneously, input autism data is normalized using Min–Max normalization, and then Matusita and Chord Distance is deployed for feature selection. Thereafter, the Synthetic Minority Oversampling Technique (SMOTE) is considered for data augmentation. The results from the above two processes are employed for detecting ASD by considering CLeFTNet. The analytic metric, namely accuracy, sensitivity and specificity attained 91.723%, 91.396% and 91.573%.

Accurate classification of brain tumors from medical images is essential for enabling timely diagnosis and effective treatment. This study aimed to develop an innovative method for the diagnosis of brain tumors through a Fuzzy Richards Functions-based Ensemble Network (FRE-Net). The parameters of the Richards function are optimized through Grid Search (GS) for selecting an optimal set of parameters. Our proposed method integrates three well-established pre-trained Convolutional Neural Networks (CNNs): MobileNetV1, MobileNetV2, ResNet50V2. To increase the robustness of these models, we incorporate a novel Lightweight Multiscale with Squeeze and Excitation (LiteMSSE) Block, which improves performance by enhancing multi-scale feature extraction and enabling the network to capture more detailed spatial information for focusing on the most relevant features to improve overall diagnostic performance. Additionally, probabilities from the individual models are aggregated using a Fuzzy Richards Functions approach, which reduces the error between observed and ground truth data, further enhancing detection accuracy. The key innovation of this study lies in the design of novel LiteMSSE Block and use of Fuzzy Richard Function, which together enhance multi-scale feature extraction and combines diverse model predictions intelligently. The proposed FRE-Net method achieves an impressive accuracy of 98.47% on the four-class Kaggle dataset and 99.00% on the BR35H dataset by highlighting its potential as a powerful tool for diagnosis of brain MRI more precisely. Through extensive evaluations, we determine that our proposed ensemble method outperforms individual backbone models and existing methods.

This study presents a novel hybrid optimization model that combines the complementary aspects of Sand Cat Swarm Optimization (SCSO) and Whale Optimization Algorithm (WOA) to solve the multi-level image thresholding problem. The proposed approach utilizes an adaptive two-stage mechanism that balances the high exploration capacity of SCSO with the concentrated local search capability of WOA, aiming to maximize inter-class variance in the histogram-based thresholding process. Various experiments are conducted on lung cancer, prostate, and mixed medical image datasets. Results demonstrate that modified SCSOWOA achieves superior performance across all datasets. For LC25000, it attains PSNR 27.9453 dB, SSIM 0.9340, FSIM 0.9542, Dice coefficient 0.8901, and Jaccard index 0.8034 at T?=?12. For prostate, PSNR reaches 28.3965 dB, SSIM 0.7532, FSIM 0.8170, Dice 0.9215, and Jaccard 0.8593. In the MSD dataset, SCSOWOA achieves PSNR 29.3244 dB, SSIM 0.7118, FSIM 0.7562, Dice 0.8901, and Jaccard 0.8034, indicating consistent performance across diverse organs and modalities. The method also demonstrates high computational efficiency, with an average execution time of 1.3221 s, offering up to 40% speed improvement over conventional metaheuristics such as PSO and GWO. Overall, proposed method provides high-accuracy, low-variance, and computationally efficient segmentation, preserving both structural and perceptual fidelity. These results confirm the method’s robustness, generalizability, and practical applicability for AI-assisted diagnostic systems across histopathological and medical imaging contexts, balancing precision, structural preservation, and speed for real-world clinical deployment.