Biomimetics

A distinct typing interface using surface electromyography (sEMG) can facilitate silent, hands-free typing by interpreting muscle activity in relation to specific keystrokes. Character-level recognition poses greater challenges than coarse gesture recognition because it is sensitive to subtle temporal variations and overlapping muscle dynamics. Temporal features are essential for typing recognition because keypresses may differ in duration, force, and accompanying hand movements across users. This paper proposes TransTCNet, a two-stage deep neural network architecture with a causal convolutional layer for learning local features and a transformer-based component for learning long-range temporal interactions. We evaluated our network on a publicly available 26-class typing sEMG dataset acquired from 19 individuals. The model achieved a validation accuracy of 96.53%, exceeding the baseline models. Our study revealed generalization among participants, and the AUC values were also high (>0.994) across all classes. The model was highly reliable and exhibited high prediction confidence (>0.9), enabling us to achieve a high training accuracy (97.86%) for real-time filtering decisions. TransTCNet could be suitable for wearable and edge devices due to its efficient architecture and low inference cost. The model’s ability to consistently decode fine-grained neuromuscular signals across users makes it well-suited for real-time applications such as adaptive user interfaces, virtual and augmented reality, prosthetic control, and communication systems. Full article

Erosive wear in pipe elbows subjected to liquid–solid two-phase flow is a major cause of material degradation and service failure in industrial piping systems. In this study, erosion characteristics of pipe elbows were investigated through erosion mapping experiments and numerical simulations. The effects of flow velocity and particle diameter on erosion location and intensity were analyzed. Erosion was found to be mainly concentrated on the outer wall of the elbow within the angular range of 10° to 90°, and both erosion intensity and affected area increased with increasing particle diameter and flow velocity. Dean vortices were shown to play an important role in particle transport and erosion distribution, especially for small particles. Inspired by the ribbed morphology of shells, a biomimetic elbow was further designed and evaluated through an orthogonal numerical study considering flow velocity, particle diameter, rib number, and rib diameter. The results indicate that the ribbed structure can effectively improve erosion resistance by altering particle trajectories, reducing particle impact probability, and dissipating kinetic energy through low-velocity rotating flow between adjacent ribs. This finding provides useful inspiration for addressing erosive wear problems in engineering applications. Full article

This paper proposes a multi-strategy improved pied kingfisher optimizer (MSIPKO), a novel metaheuristic algorithm designed to address constrained optimization problems (COPs). COPs are widely encountered in engineering and industrial applications and are characterized by complex constraints that restrict the feasible solution space and often lead to multiple local optima. To enhance the performance of the original pied kingfisher optimizer (PKO), three strategies are incorporated: (i) a reverse differential crossover mechanism to improve global exploration and maintain population diversity; (ii) an enhanced diving-fishing operator to strengthen local exploitation; and (iii) an improved commensalism phase to enrich search directions and increase robustness. The performance of MSIPKO is evaluated on 12 benchmark functions from the IEEE Congress on Evolutionary Computation 2006 (CEC 2006) test suite and six classical engineering optimization problems. Experimental results demonstrate that MSIPKO outperforms several state-of-the-art algorithms in terms of optimization accuracy, convergence speed, and stability, particularly for high-dimensional, nonlinear, and multi-constrained problems. Moreover, MSIPKO achieves superior or comparable solutions with fewer function evaluations, indicating its high efficiency and adaptability. These results confirm that MSIPKO is a promising tool for solving complex real-world constrained optimization problems. Future work will focus on extending the proposed algorithm to multi-objective and large-scale optimization scenarios. Full article

Ischemic stroke, a major cause of global disability, is characterized by the blockage of an artery leading to reduced cerebral blood flow and subsequent brain injury. Automatic segmentation of ischemic stroke lesions in Computed Tomography Perfusion (CTP) maps is critical for accurate diagnosis, treatment planning, and outcome assessment. However, the accuracy of traditional methods remains limited, with Dice Similarity Coefficient (DSC) values around 68%. To address this challenge, we propose a deep learning-based model inspired by biological systems and brain mechanisms, which emulates natural information processing to enhance ischemic stroke lesion segmentation. The proposed network architecture consists of five graph convolutional layers that automatically extract and classify features from CTP images. We evaluated the model using the ISLES 2018 database, achieving a DSC of 75.41% and a Jaccard Index of 74.52%, representing significant improvements over previous methods. Notably, the proposed approach performs robustly in noisy environments, maintaining accuracy above 60% even at SNR = −4. These results demonstrate the potential of biomimetic-inspired networks for automatic ischemic stroke segmentation. Full article

Currently, small moving object detection remains a highly challenging problem, primarily attributable to four critical factors: limited pixel coverage, blurred texture features, indistinguishability from small-object-like background features (i.e., false positives), and vulnerability to environmental noise interference. The remarkable sensitivity of the Drosophila visual system to small moving objects, which originates from a specialized type of neuron known as “lobula columnar 11” (LC11), has provided inspiration for addressing this challenge. Current bio-inspired visual models have achieved certain advances. However, detection performance against real-world complex dynamic natural environments still requires further improvement. To address the challenge of limited detection accuracy for small moving objects against real-world complex dynamic natural environments, this paper proposes a Motion Small Object Detection (MSOD) model inspired by the Drosophila Vision Small Object Motion Sensitivity (DVSOMS) mechanism, namely DVSOMS-MSOD. The model consists of four stages: The first stage is preliminary processing of visual stimuli, where visual stimuli are perceived, converted to grayscale, and blurred. The second stage is the motion neural pathway, where visual signals are first decomposed into parallel ON and OFF neural pathway signals; then, the neural feedback mechanism is incorporated between the medulla and lobula complex, and the complete Hassenstein–Reichardt correlator (HRC) is integrated into the lobula complex; finally, the LC11 neuron is utilized to detect small moving objects and extract their location information. The third stage is the contrast neural pathway, where visual signals are first processed by the central and surrounding local neighborhoods, then local contrast information is calculated. The fourth stage is the integration of motion and contrast neural pathways, where the mushroom body generates motion trajectories using the location information of small moving objects, and subsequently generates contrast trajectories using the local contrast information and motion trajectories to more finely detect small moving objects. Under real-world complex dynamic natural environment datasets, compared with conventional machine learning methods for moving object detection, the proposed model achieved improvements of 77.82% and 78.70% in detection performance and output quality, respectively, while reducing running time by 10.60%. Compared with bio-inspired visual models for small moving object detection, the proposed model achieved improvements of 28.24% and 43.15% in detection accuracy and detection performance, respectively, but the running time increased by 43.40%. The proposed model demonstrates certain advantages in detection performance, output quality, and detection accuracy; however, its real-time performance still warrants further optimization. Full article

Tree-climbing robots are primarily utilized for pruning and harvesting in tall trees; however, limited structural degrees of freedom (DoFs) reduce their flexibility in complex environments. To improve the flexibility and environmental adaptability of the robots, this study proposes a novel three-armed claw-type tree-climbing robot inspired by gibbons. A 14 DoFs prototype with a total mass of approximately 2.52 kg was developed, comprising three manipulator arms and independently actuated claws. Kinematic models were separately established for the series-connected arms and the parallel-connected moving platform, with accuracy verified through numerical simulations. Based on these models, a control system was implemented, and a physical prototype was tested in field climbing experiments. Grasping tests on surfaces of varying roughness, including moist tree trunks, artificial wood, and smooth steel plates, demonstrated the adaptability of the claw to diverse materials. The robot successfully climbed trunks inclined at 52–90°, supporting a maximum payload of 1.81 kg; each full gait cycle averaged approximately 4 min. These results indicate that the robot can successfully imitate the movements of gibbons during climbing, thereby verifying the feasibility and practical application value of this bionic design in real-world forestry environments. Full article

This study addresses pressure regulation in an induction-motor-driven centrifugal fan and introduces two complementary novelties: a Softsign-PI controller that shapes the tracking error via a Softsign nonlinearity before PI regulation and a hybrid starfish optimization with a differential evolution (hSFOA-DE) scheme for automatically tuning the controller parameters. The approach is evaluated on an experimentally validated nonlinear fan–motor model and benchmarked against modern metaheuristics—starfish optimization algorithm (SFOA), animated oat optimization (AOO), electric eel foraging optimization (EEFO), differential evolution (DE), particle swarm optimization (PSO)—as well as classical tunings—Murrill-based 2-DOF PID, Tyreus–Luyben PID and Ziegler–Nichols PI. Statistical summaries and boxplots indicate superior central tendency with reduced run-to-run variability; fitness–evolution curves show faster convergence; and time-domain performance metrics confirm improved transient and steady-state behaviour. Objective function comparisons further show the lowest values of both the Zwe-Lee Gaing (ZLG) and integral of absolute error (IAE), supporting advantages in robustness and tracking accuracy of the proposed approach. These gains reduce overshoot and cumulative error, which can lessen throttling losses and actuator duty in fan/pump service, suggesting potential energy and maintenance benefits. Full article

Biomaterials play a central role in tissue engineering and regeneration by providing scaffolds that support cell adhesion, proliferation and differentiation while modulating the surrounding microenvironment. They represent promising alternatives to traditional surgical approaches that may lead to complications or tissue damage, and their performance is influenced by chemical composition, mechanical behavior, architecture and interfacial properties, all of which can be precisely tuned through advanced fabrication and surface modification strategies. This review synthesizes evidence from a comprehensive literature search across major scientific databases, focusing on highly cited studies and available clinical data, and examines natural and synthetic biomaterials, their biological responses, functional characteristics, and surface modification methods. Emphasis is placed on Cold Atmospheric Plasma (CAP), which selectively modifies the outermost nanolayer of materials, enhancing hydrophilicity, functional group density, protein adsorption and overall cell–material interactions, as well as improving drug loading capacity. The review also considers stem cell interactions with biomaterials and emerging applications of artificial intelligence (AI) for predicting performance and guiding material optimization. Overall, the analysis highlights that natural matrices provide intrinsic bioactivity, synthetic polymers offer tunable mechanics and degradation profiles, and composite systems integrate these advantages. Advances in technologies such as electrospinning and 3D/4D printing enable precise control over architecture, supporting cell colonization and vascularization. Collectively, developments in CAP treatments and AI-driven design strategies are strengthening the regenerative potential of biomaterials and advancing their clinical translation. Full article

Nature-inspired material design has gained increasing attention in the development of sustainable biocomposites for applications requiring the integration of structural performance and functional efficiency. However, many lignocellulosic composites still depend on synthetic binders and fail to achieve a strong effective interaction between constituents, resulting in suboptimal mechanical integrity and thermal behavior while limiting their environmental advantages. This study aims to develop binderless biocomposite panels from Hevea brasiliensis-derived spent mushroom substrate (SMS) and Elaeis guineensis empty fruit bunch (EFB) fibers, emphasizing the synergistic interaction between components for energy-efficient building applications. Chemical characterization revealed complementary roles, with EFB contributing a high cellulose content (57.60%) for reinforcement and SMS providing a higher lignin content (30.51%) for enhanced rigidity and natural binding. Panels were fabricated via hot pressing at a target density of 0.8 g/cm³ without additives. Mechanical properties were evaluated through specific flexural, tensile, internal bond, and impact testing, while thermal conductivity and thickness swelling were used to assess functional performance. The 60% SMS with 40% EFB composition exhibited optimal performance, achieving a specific flexural strength of 20.26 MPa, a flexural modulus of 1943.76 MPa, tensile strength of 6.12 MPa, an internal bond strength of 2.06 MPa, an impact strength of 15.35 kJ/m², a thickness swelling of 44.80%, and a thermal conductivity of 0.234 W/m.K. These results demonstrate that the combined effect of SMS and EFB in binderless biocomposites derived from secondary products offers a promising biomimetic pathway for designing recyclable, high-performance materials suitable for sustainable and energy-efficient construction systems. Full article

In this study, a novel flying legged robot configuration with enhanced obstacle-crossing capability is introduced. Legged robots, especially RHex robots, already possess high obstacle-crossing capability; however, the obstacle size that can be overcome is directly dependent on the leg length. Although stair climbing–descending, obstacle course and inclined surface algorithms have been studied for the RHex robot, flight capability has not been explored. In this study, this improvement is achieved with minimal impact on the RHex’s design by adding just a thruster as an additional propulsion system to propel the robot into flight. The attitude control is realized using the mass actuation of the robot legs, similar to how animals like lizards and cats utilize their limbs or tails as inertial appendages to stabilize their body pitch during mid-air maneuvers. Instead of direct and complete flight control, the aim was a temporary flight similar to obstacle-clearing flights of chickens. Hence, a nonlinear 2D model is developed to investigate the kinematics and dynamics of the RHex robot. Equations of motion are derived, linearized and used in a state feedback regulator design; the regulator is also expanded for reference tracking. Full article

This study reports on the physico-chemical and in vitro biological characterization of chromium-doped hydroxyapatite (10CrHAp, Cr³⁺, Ca_10-xCr_x(PO₄)₆(OH)₂, x_Cr = 0.1) and chromium-doped hydroxyapatite in dextran matrix (10CrHAp-Dx) coatings, prepared for the first time via the spin coating technique. X-ray diffraction analysis and Rietveld refinement were used to characterize the materials. Fourier-transform infrared (FTIR) spectroscopy confirmed the presence of functional groups specific to hydroxyapatite. Scanning electron microscopy (SEM) observations revealed the presence of a conglomerate of nanoparticles distributed unevenly across the coatings surface. Atomic force microscopy (AFM) showed that both coatings presented continuous surfaces with uniform morphology. The in vitro biocompatibility of 10CrHAp and 10CrHAp-Dx coatings was evaluated using human osteoblast-like MG63 cell line and MTT assay. SEM and MM visualization assessed the cell adhesion and proliferation and morphological changes in the adhered cells. The antibacterial properties of the 10CrHAp and 10CrHAp-Dx coatings was assessed in vitro against two of the most common bacterial reference strains, Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 25923. Overall, the coatings achieved log reductions up to ~9.35, corresponding to a bacterial kill rate (for S. aureus) exceeding 99.99%, with 10CrHAp-Dx showing slightly superior performance. Similar behavior (log reductions of ~8.6 and ~8.9, respectively, indicating a sustained antibacterial effect and >99.99% bacterial elimination) was observed and for Pseudomonas aeruginosa. AFM was used to evaluate the bacterial cells interaction with the coating’s surfaces. The biological assays demonstrated that both coatings possess notable antibacterial activity, underscoring their potential in biomedical applications, particularly in the design of new antimicrobial devices. Full article

Low-altitude terrain-following flight is essential for obtaining high-quality data in unmanned aerial vehicle (UAV) aeromagnetic surveys, but achieving efficient and safe path planning within complex terrains remains challenging. To address this issue, a Q-learning-assisted multi-strategy Starfish Optimization Algorithm (QMSFOA) is proposed for offline path planning. The proposed algorithm integrates four improvement strategies: (1) employing a Sobol sequence combined with Refraction Opposition-based Learning for population initialization to enhance population diversity; (2) adopting a hybrid adaptive differential mutation mechanism to improve search efficiency; (3) utilizing Q-learning to intelligently schedule optimization modes, thereby accelerating convergence speed; (4) introducing an adaptive t-distribution elite perturbation strategy to refine convergence accuracy. Experimental results on the CEC-2022 benchmark suite indicate that QMSFOA achieves the best convergence accuracy on nine functions and exhibits a superior performance across most metrics compared with the competing algorithms. Simulation experiments of aeromagnetic surveys in complex 3D terrains demonstrate that paths planned by QMSFOA satisfy kinematic and obstacle avoidance constraints while reducing path costs by approximately 25% compared with the standard Starfish Optimization Algorithm (SFOA). Additionally, the standard deviation is reduced by one to two orders of magnitude compared with the competing algorithms. These results demonstrate that the proposed method provides an efficient, reliable, and intelligent solution for high-precision UAV geophysical exploration in complex environments. Full article

Ocean monitoring is essential for understanding climate change and marine ecosystem dynamics, yet achieving comprehensive global coverage remains a challenge in oceanography. Current technologies face limitations in cost, power, hardware, and depth capacity that restrict widespread monitoring capabilities. Here we show that biohybrid robotic jellyfish (Aurelia aurita) can serve as autonomous vertical ocean profilers by integrating microcontrollers with positively buoyant sensor payloads, achieving controlled vertical-profiling capabilities. Laboratory experiments demonstrated repeatable up–down trajectories, quantified force balance limits, and identified predictable, size-dependent descent swimming speeds. Field deployments in Massachusetts coastal waters and the open ocean off the Florida Keys demonstrated field operation to ocean depths >25 m with successful in situ temperature and depth measurements. To our knowledge, this represents the first biohybrid jellyfish platform to combine autonomous, pressure-triggered vertical profiling with onboard oceanographic sensing in natural marine environments. This approach leverages the global distribution and remarkable swimming efficiency of living jellyfish while eliminating propulsion power requirements by utilizing the animal’s natural swimming capabilities. While further development is required for long-term ocean deployment, this study lays the groundwork for a new class of biohybrid ocean-sensing platforms with advantages in cost, power, and mission flexibility, providing a pathway toward dense sensor networks and increased ocean monitoring observations. Full article

In organizational cyber crises, incident response and official communication form coupled control loops, yet they are usually engineered separately. We present V-CHIMERA (Verified Coupled Human–Information–Machine Incident Response Architecture), a framework for organizational cyber crisis response under misinformation that jointly models cyber state, belief dynamics, trust, and communication governance. The framework combines three elements: an explicit cyber–social coupling architecture, a runtime protocol shield for communication safety, and immune-gated coupling (IGC) that uses danger signaling, tolerance thresholds, and immune memory to regulate when social feedback should affect operational response and how strongly counter-messaging should be targeted. Across three representative scenarios—ransomware rumor, outage rumor, and exfiltration scam—and eight seeds per scenario, all shielded policies achieved zero executed protocol violations. Relative to naive coupled control, IGC reduced cyber-harm area under the curve (AUC) by 57.6% in ransomware rumor and 42.6% in outage rumor while also reducing misbelief. Results were scenario-dependent rather than uniformly dominant: in exfiltration scam, a broadcast-only ablation outperformed targeted messaging, showing that targeting can fail when diffusion rapidly crosses community boundaries. Sensitivity analysis further shows that IGC attenuates the brittleness observed under strong coupling and weak moderation. The results suggest that biomimetic regulation is valuable not because coupling always helps, but because it prevents overreaction, clarifies when targeting should be used, and yields safer organizational defaults for misinformation-aware incident response. Full article

Stress urinary incontinence (SUI) affects a significant proportion of women and often requires surgical intervention when conservative treatments fail. While midurethral slings (MUS) are widely used, concerns over complications such as mesh exposure/erosion and chronic pain have driven interest in regenerative medicine alternatives. This review explores emerging strategies, including stem cell therapies, platelet-rich plasma injections, decellularized extracellular matrix scaffolds, injectable hydrogels, and bioengineered slings. These approaches aim to restore continence by promoting tissue regeneration, improving biocompatibility, and reducing adverse reactions. We evaluate their mechanisms, reported outcomes, and current stage of development, supported by in vitro and in vivo model data. Although promising, these technologies face challenges related to cell viability, scaffold integration, and clinical translation. Continued interdisciplinary research is essential to optimize these therapies and bring safer, more effective solutions to patients. Regenerative strategies may ultimately redefine the future of SUI treatment by offering biologically integrated, long-lasting alternatives to synthetic slings. To date, no tissue-engineered or regenerative biomimetic sling has received regulatory approval for routine clinical use in the management of stress urinary incontinence. Full article

Metaheuristic algorithms can fail to balance global exploration and local exploitation, occasionally becoming trapped in suboptimal regions on highly multimodal problems. The Groupers and Moray Eels (GME) algorithm, inspired by the associative hunting strategies of marine predators, provides a cooperative optimization framework. However, the sequential interaction phases of GME can fail to maintain diverse topological coverage across heavily constrained landscapes. To address these limitations, we propose an enhanced variant, GPS-CC-GME. The approach improves the initial agent distribution by deploying a number-theoretic Good Point Set (GPS) generation protocol to establish a uniformly dispersed starting space. In addition, algorithmic stagnation is addressed through a dual-crossover search architecture. A horizontal crossover stage enforces information sharing among randomized agents to sustain global diversity, and a vertical crossover phase isolates specific dimensional vectors within individual agents for localized fine-tuning. We evaluated the proposed model on the CEC2017 benchmark suite, where it secured the highest overall ranking compared to the baseline GME and several standard metaheuristics. GPS-CC-GME was then applied to a high-dimensional optimization scenario for petroleum reservoir production. The algorithm yielded higher Net Present Value (NPV) metrics than the canonical framework. The results indicate that embedding deterministic initialization and bidirectional mutation operators into multipredator models can improve search outcomes in non-linear engineering tasks. Full article

Aiming at the problems of uneven population initialization distribution, easy trapping in local optima, unbalanced exploration and exploitation capabilities, insufficient optimization accuracy and convergence speed of the original Greater Cane Rat Algorithm (GCRA), this paper proposes a Chaos-Integrated Difference-Enhanced Greater Cane Rat Algorithm (CEGCRA). Firstly, the algorithm adopts the piecewise chaotic map to generate the initial population, which effectively improves the uniformity and diversity of the population and reduces the risk of premature convergence. Secondly, an accumulated difference foraging strategy is designed to integrate the position and fitness difference information between individuals and the optimal individual, dynamically adjust the search direction and step size, and realize the adaptive balance between global exploration and local exploitation capabilities. Finally, the dynamic switching mechanism between the exploration and exploitation stages of the algorithm is improved, and the boundary constraint handling strategy is optimized to further enhance the algorithm stability. To verify the performance of the CEGCRA, comparative experiments were carried out on the CEC2014 and CEC2020 benchmark test suites. The results show that compared with the original GCRA, the optimal fitness value of the CEGCRA is reduced by an average of 35.3%, the standard deviation is reduced by an average of 22.7%, and the convergence speed is increased by an average of 28.9%. In two typical engineering constrained optimization problems, namely, welded beam design and cantilever beam design, the cost of the welded beam solved by the CEGCRA is 12.5% lower than that of the original GCRA and 8.7% lower than that of the PSO algorithm; the weight of the cantilever beam is 0.012% lower than that of the original GCRA and 0.008% lower than that of the GA, with a constraint satisfaction rate of 100%. The experimental results fully prove that the CEGCRA is superior to the original GCRA and seven comparison algorithms such as PSO, DE and SSA in terms of optimization accuracy, convergence speed, robustness and constraint handling ability and can effectively solve complex engineering optimization problems with high dimensionality, nonlinearity and multiple constraints. Full article

Biomimetic materials mimic biological structures and functions. They are crucial for addressing complex challenges in tissue engineering, sustainable architecture, and energy storage. Traditionally, designing these materials requires slow, resource-intensive trial-and-error methods and physics-based simulations. Recently, Artificial Intelligence (AI) and Machine Learning (ML) have transformed this field. They translate biological intelligence into actionable engineering logic and rapidly explore massive design spaces. Despite rapid advancements, the field still faces several critical bottlenecks, including complexity mismatches, data scarcity, and limited interpretability. This review examines AI-driven biomimetic design across five primary “interfaces”: (1) Biological Pattern Recognition, (2) Structural Optimization, (3) Generative Morphogenesis, (4) Adaptive Fabrication, and (5) Data-Driven Discovery Platforms. The review also outlines future perspectives, especially the shift toward autonomous “closed-loop” laboratories. In these labs, AI will manage the entire workflow, i.e., design, synthesis, and testing, without human intervention. Future efforts will likely focus on multi-model data mining to understand complex, life-like properties. Furthermore, research aims to develop Explainable AI (XAI) to ensure deterministic modeling in safety-critical applications. The ultimate goal is a synergistic relationship. AI will design materials, but these materials, using biomimetic metabolic or neural models, will also help construct more efficient AI architectures. Full article

Data-driven exoskeletons offer the potential for adaptive augmentation of human mobility. Yet their widespread adoption is hindered by labor-intensive biomechanical data collection and manual tuning. Herein, this study presents a highly efficient synthetic data approach to facilitate data-driven pipelines. We leveraged an Adversarial Motion Priors (AMP) agent to learn stylized walking within a massively parallel, physics-based simulation. The resulting high-fidelity data were collected and validated against OpenSim inverse dynamics pipelines. Further, we trained an end-to-end torque prediction algorithm using the collected data. A novel CNN-Transformer architecture was developed to map contralateral swing-phase data to variable-length push-off torque profiles. This enabled real-time, adaptive torque assistance of exoskeletons for variable-speed walking. A custom ankle exoskeleton was used to demonstrate robust sim-to-real transferability. Our system achieved an average root mean square error of approximately 0.081 ± 0.015 newton-meters per kilogram and an average R² of 0.836 ± 0.050 across speeds ranging from 0.6 to 1.75 m·s⁻¹. The controller significantly reduced user-positive ankle mechanical work by up to 14 ± 6.30%. Finally, our multi-sensor configuration exhibited inherent fault tolerance, ensuring safe operation even under partial sensor failure. By taking a scalable, data-driven approach, this work offers a practical pathway toward deploying autonomous exoskeletons in versatile, real-world environments. Full article