Biomimetics最新文献

The Pathfinder Algorithm (PFA) is a bionic swarm intelligence optimization algorithm inspired by simulating the cooperative movement of animal groups in nature to search for prey. Based on fitness, the algorithm classifies search individuals into leaders and followers. However, PFA fails to effectively balance the optimization capabilities of leaders and followers, leading to problems such as insufficient population diversity and slow convergence speed in the original algorithm. To address these issues, this paper proposes an enhanced pathfinder algorithm based on multi-strategy (EODE-PFA). Through the synergistic effects of multiple improved strategies, it effectively solves the balance problem between global exploration and local optimization of the algorithm. To verify the performance of EODE-PFA, this paper applies it to CEC2022 benchmark functions, three types of complex engineering optimization problems, and six sets of feature selection problems, respectively, and compares it with eight mature optimization algorithms. Experimental results show that in three different scenarios, EODE-PFA has significant advantages and competitiveness in both convergence speed and solution accuracy, fully verifying its engineering practicality and scenario universality. To highlight the synergistic effects and overall gains of multiple improved strategies, ablation experiments are conducted on key strategies. To further verify the statistical significance of the experimental results, the Wilcoxon signed-rank test is performed in this study. In addition, for feature selection problems, this study selects UCI real datasets with different real-world scenarios and dimensions, and the results show that the algorithm can still effectively balance exploration and exploitation capabilities in discrete scenarios.

This study proposes a Hybrid Strategy Improved Dingo Optimization Algorithm (HSIDOA), designed to address the limitations of the standard DOA in complex optimization tasks, including its tendency to fall into local optima, slow convergence speed, and inefficient boundary search. The HSIDOA integrates a quadratic interpolation search strategy, a horizontal crossover search strategy, and a centroid-based opposition learning boundary-handling mechanism. By enhancing local exploitation, global exploration, and out-of-bounds correction, the algorithm forms an optimization framework that excels in convergence accuracy, speed, and stability. On the CEC2017 (30-dimensional) and CEC2022 (10/20-dimensional) benchmark suites, the HSIDOA achieves significantly superior performance in terms of average fitness, standard deviation, convergence rate, and Friedman test rankings, outperforming seven mainstream algorithms including MLPSO, MELGWO, MHWOA, ALA, HO, RIME, and DOA. The results demonstrate strong robustness and scalability across different dimensional settings. Furthermore, HSIDOA is applied to multi-level threshold image segmentation, where Otsu's maximum between-class variance is used as the objective function, and PSNR, SSIM, and FSIM serve as evaluation metrics. Experimental results show that HSIDOA consistently achieves the best segmentation quality across four threshold levels (4, 6, 8, and 10 levels). Its convergence curves exhibit rapid decline and early stabilization, with stability surpassing all comparison algorithms. In summary, HSIDOA delivers comprehensive improvements in global exploration capability, local exploitation precision, convergence speed, and high-dimensional robustness. It provides an efficient, stable, and versatile optimization method suitable for both complex numerical optimization and image segmentation tasks.

In response to the challenges posed by high operational resistance and significant soil adhesion faced by traditional pressing rollers in moist clay soils, this study introduces a bionic pressing roller inspired by the imbricated scale structure of the pangolin. The fundamental working principles of the roller are elucidated, and its key structural parameters are designed. Utilizing the discrete element method (DEM), the structural parameters of the bionic scales are optimized through Response Surface Methodology (RSM), with traveling resistance and the mass of adhered soil serving as evaluation indicators. Field experiments are conducted to validate the operational performance of the bionic roller. The optimal parameter combination is identified as follows: a scale length of 130 mm, 10 scales, and an overlap rate of 50%. Field comparison tests reveal that the bionic roller significantly reduces traveling resistance by 11.0% and decreases the mass of adhered soil by 47.2% compared to the traditional roller at a soil moisture content of 35%. This study confirms that the bionic roller, which mimics the pangolin scale structure, demonstrates superior anti-adhesion and drag-reduction characteristics. The findings are anticipated to provide a reference for the energy-efficient design of soil-engaging components in agricultural machinery, including ridging and shaping machines.

To address the issue of drought level confusion in the detection of drought stress during the seedling stage of the Yunnan large-leaf tea variety using the traditional YOLOv13 network, this study proposes an improved version of the network, MC-YOLOv13-L, based on animal vision. With the compound eye's parallel sampling mechanism at its core, Compound-Eye Apposition Concatenation optimization is applied in both the training and inference stages. Simulating the environmental information acquisition and integration mechanism of primates' "multi-scale parallelism-global modulation-long-range integration," multi-scale linear attention is used to optimize the network. Simulating the retinal wide-field lateral inhibition and cortical selective convergence mechanisms, CMUNeXt is used to optimize the network's backbone. To further improve the localization accuracy of drought stress detection and accelerate model convergence, a dynamic attention process simulating peripheral search, saccadic focus, and central fovea refinement in primates is used. Inner-IoU is applied for targeted improvement of the loss function. The testing results from the drought stress dataset (324 original images, 4212 images after data augmentation) indicate that, in the training set, the Box Loss, Cls Loss, and DFL Loss of the MC-YOLOv13-L network decreased by 5.08%, 3.13%, and 4.85%, respectively, compared to the YOLOv13 network. In the validation set, these losses decreased by 2.82%, 7.32%, and 3.51%, respectively. On the whole, the improved MC-YOLOv13-L improves the accuracy, recall rate and mAP@50 by 4.64%, 6.93% and 4.2%, respectively, on the basis of only sacrificing 0.63 FPS. External validation results from the Laobanzhang base in Xishuangbanna, Yunnan Province, indicate that the MC-YOLOv13-L network can quickly and accurately capture the drought stress response of tea plants under mild drought conditions. This lays a solid foundation for the intelligence-driven development of the tea production sector and, to some extent, promotes the application of bio-inspired computing in complex ecosystems.

To avoid predators, benthic fish will stir up the sediment on the seabed by flapping their pectoral fins, thus burying themselves. This self-burial concealment strategy can offer bionic enlightenment for the benthic residence method of Unmanned Underwater Vehicles (UUVs). In this paper, based on the observation results of the self-burial behavior of benthic fish, a two-dimensional fluid-particle numerical model was developed to simulate the processes of sediment transport induced by pectoral fin flapping. In addition, an orthogonal experimental design was employed to analyze the effects of body length, flapping amplitude, flapping number, flapping frequency, and particle size on burial ratio, input power, and burial efficiency. The results reveal that rapid pectoral fin flapping enables benthic fish to fluidize sediments and achieve self-burial. Among the influencing factors, body size has the most significant impact on coverage ratio and input power, as larger fish generate stronger tip vortices and fluid disturbances, making local flow velocities more likely to exceed the critical starting velocity. In contrast, particle size has the weakest effect on burial performance, while kinematic parameters exert a far greater impact on self-burial than environmental parameters. The research results can offer references for the biomimetic design of self-burying UUVs.

Artificial Intelligence (AI) has seen rapid advancements in recent times, finding applications across various sectors and achieving notable successes. However, current AI models based on Deep Convolutional Neural Networks (DNNs) face numerous challenges, particularly a lack of interpretability, which severely restricts their future potential. Spiking Neural Networks (SNNs), considered the third generation of Artificial Neural Networks (ANNs), are at the forefront of brain-inspired AI research. The resemblance between SNNs and biological neural networks offers the potential to create more human-like AI systems with enhanced interpretability, paving the way for more trustworthy AI implementations. Despite this promise, the absence of efficient training methods for large-scale and complex SNNs hampers their broader application. This paper investigates bio-inspired reinforcement learning strategies by examining neural network dynamics during SNN training. The aim is to improve learning efficiency and effectiveness for extensive and intricate SNNs. Our findings suggest that using reinforcement learning to focus on neural network dynamics may be a promising approach for developing learning algorithms for future large-scale SNNs.

Jumping is a fundamental locomotion in insects, offering high performance and efficient movement. However, the relationships between the jumping force and performance remain inadequately understood. Here, we combine experimental measurements with a theoretical model to investigate the jumping kinematics and performance of crickets Velarifictorus micado. We examine how jumping force, gravity, aerodynamic drag, and take-off angle influence the jumping velocity, displacement, and power output of the crickets. We discuss the mechanistic advantages of various jumping force designs and demonstrate that the front slow-loaded force adopted by crickets enables greater power output while minimizing take-off displacement and acceleration time. The results show that aerodynamic drag exerts negligible influence, whereas gravity mainly affects the vertical propulsive component during the take-off phase. The gravitational effect leads to a decrease in resultant velocity and displacement with increasing take-off angle. This study advances our understanding of the mechanical principles governing jumps of insects and provides valuable insights for the design of high-performance jumping robots and catapult systems.

Contemporary neural and generative architectures are deficient in self-preservation mechanisms and sustainable stability. In uncertain or noisy situations, they frequently demonstrate oscillatory learning, overconfidence, and structural deterioration, indicating a lack of biological regulatory principles in artificial systems. We present Bio-RegNet, a meta-homeostatic Bayesian neural network architecture that integrates T-regulatory-cell-inspired immunoregulation with autophagic structural optimization. The model integrates three synergistic subsystems: the Bayesian Effector Network (BEN) for uncertainty-aware inference, the Regulatory Immune Network (RIN) for Lyapunov-based inhibitory control, and the Autophagic Optimization Engine (AOE) for energy-efficient regeneration, thereby establishing a closed energy-entropy loop that attains adaptive equilibrium among cognition, regulation, and metabolism. This triadic feedback achieves meta-homeostasis, transforming learning into a process of ongoing self-stabilization instead of static optimization. Bio-RegNet routinely outperforms state-of-the-art dynamic GNNs across twelve neuronal, molecular, and macro-scale benchmarks, enhancing calibration and energy efficiency by over 20% and expediting recovery from perturbations by 14%. Its domain-invariant equilibrium facilitates seamless transfer between biological and manufactured systems, exemplifying a fundamental notion of bio-inspired, self-sustaining intelligence-connecting generative AI and biomimetic design for sustainable, living computation. Bio-RegNet consistently outperforms the strongest baseline HGNN-ODE, improving ARI from 0.77 to 0.81 and NMI from 0.84 to 0.87, while increasing equilibrium coherence κ from 0.86 to 0.93.

Energy-absorbing components should be effective and stable in engineering protective structure designs to reduce collision impacts. However, conventional energy-absorbing structures have considerable potential for optimization for energy dissipation and structural stability. Like other invertebrates, the centipede's folding mode when moving forward is compatible with the hierarchical folding process when the energy-absorbing structure is impacted; however, this rule has not been thoroughly examined and proven. Based on this gap, this study built a unique biomimetic aluminum foam-filled bidirectional pyramid energy-absorbing structure, analyzed its geometric parameters on crashworthiness, and developed high-performance energy-absorbing components. Experiments and simulations were conducted on a bidirectional pyramid construction with three schemes for filling aluminum foam inspired by the centipede body section and profile. The construction with foam aluminum filling the gap has optimum specific energy absorption and load stability. Additionally, optimizing structural performance is most effective in certain ranges (78° ≤ θ ≤ 87°, t ≤ 0.1 mm, 34 mm ≤ d ≤ 44 mm). With Kriging and NSGA-III multi-objective optimization, the optimized peak crushing force decreases by 11.17% and specific energy absorption increases by 11.67%. The study and optimization process offers a theoretical reference for future high-performance energy-absorbing structures and has significant engineering application potential.

Currently, wireless sensor networks (WSNs) have been mutually applied to environmental monitoring and industrial control due to their low-cost and low-energy sensor nodes. However, WSNs are composed of a large number of energy-limited sensor nodes, which requires balancing the relationship among energy consumption, transmission delay, and network lifetime simultaneously to avoid the formation of energy holes. In nature, gregarious herbivores, such as the white-bearded wildebeest on the African savanna, employ a "fast-transit and selective-dwell" strategy when searching for water; they cross low-value regions quickly and prolong their stay in nutrient-rich pastures, thereby minimizing energy cost while maximizing nutrient gain. Ants, meanwhile, dynamically evaluate the "energy-to-reward" ratio of a path through pheromone concentration and its evaporation rate, achieving globally optimal foraging. Inspired by these two complementary biological mechanisms, our study proposes a novel ACO-conceptualized optimization model formulated via mixedinteger linear programming (MILP). By mapping the pheromone intensity and evaporation rate into the MILP energy constraints and cost functions, the model integrates discrete decision-making (path selection) and continuous variables (dwell time) by dynamic path planning and energy optimization of mobile sink, constituting multi-objective optimization. Firstly, we can achieve flexible trade-offs between multiple objectives such as data transmission delay and energy consumption balance through adjustable weight coefficients of the MILP model. Secondly, the method transforms complex path planning and scheduling problems into deterministic optimization models with theoretical global optimality guarantees. Finally, experimental results show that the model can effectively optimize network performance, significantly improve energy efficiency, while ensuring real-time performance and extended network lifetime.