Journal of Bionic Engineering最新文献

Feature fusion is an important technique in medical image classification that can improve diagnostic accuracy by integrating complementary information from multiple sources. Recently, Deep Learning (DL) has been widely used in pulmonary disease diagnosis, such as pneumonia and tuberculosis. However, traditional feature fusion methods often suffer from feature disparity, information loss, redundancy, and increased complexity, hindering the further extension of DL algorithms. To solve this problem, we propose a Graph-Convolution Fusion Network with Self-Supervised Feature Alignment (Self-FAGCFN) to address the limitations of traditional feature fusion methods in deep learning-based medical image classification for respiratory diseases such as pneumonia and tuberculosis. The network integrates Convolutional Neural Networks (CNNs) for robust feature extraction from two-dimensional grid structures and Graph Convolutional Networks (GCNs) within a Graph Neural Network branch to capture features based on graph structure, focusing on significant node representations. Additionally, an Attention-Embedding Ensemble Block is included to capture critical features from GCN outputs. To ensure effective feature alignment between pre- and post-fusion stages, we introduce a feature alignment loss that minimizes disparities. Moreover, to address the limitations of proposed methods, such as inappropriate centroid discrepancies during feature alignment and class imbalance in the dataset, we develop a Feature-Centroid Fusion (FCF) strategy and a Multi-Level Feature-Centroid Update (MLFCU) algorithm, respectively. Extensive experiments on public datasets LungVision and Chest-Xray demonstrate that the Self-FAGCFN model significantly outperforms existing methods in diagnosing pneumonia and tuberculosis, highlighting its potential for practical medical applications.

A heart attack disrupts the normal flow of blood to the heart muscle, potentially causing severe damage or death if not treated promptly. It can lead to long-term health complications, reduce quality of life, and significantly impact daily activities and overall well-being. Despite the growing popularity of deep learning, several drawbacks persist, such as complexity and the limitation of single-model learning. In this paper, we introduce a residual learning-based feature fusion technique to achieve high accuracy in differentiating abnormal cardiac rhythms heart sound. Combining MobileNet with DenseNet201 for feature fusion leverages MobileNet lightweight, efficient architecture with DenseNet201, dense connections, resulting in enhanced feature extraction and improved model performance with reduced computational cost. To further enhance the fusion, we employed residual learning to optimize the hierarchical features of heart abnormal sounds during training. The experimental results demonstrate that the proposed fusion method achieved an accuracy of 95.67% on the benchmark PhysioNet-2016 Spectrogram dataset. To further validate the performance, we applied it to the BreakHis dataset with a magnification level of 100X. The results indicate that the model maintains robust performance on the second dataset, achieving an accuracy of 96.55%. it highlights its consistent performance, making it a suitable for various applications.

Pacinian Corpuscle (PC) is the largest tactile vibration receptor in mammalian skin, with a layered structure that enables signal amplification and high-pass filtering functions. Modern robots feature vibro-tactile sensors with excellent mechanical properties and fine resolution, but these sensors are prone to low-frequency noise interference when detecting high-frequency vibrations. In this study, a bionic PC with a longitudinally decreasing dynamic fractal structure is proposed. By creating a lumped parameter model of the PC’s layered structure, the bionic PC made of gelatin-chitosan based hydrogel can achieve high-pass filtering and specific frequency band signal amplification without requiring back-end circuits. The experimental results demonstrate that the bionic PC retains the structural characteristics of a natural PC, and the influence of structural factors, such as the number of layers in its shell, on filtration characteristics is explored. Additionally, a vibration source positioning experiment was conducted to simulate the earthquake sensing abilities of elephants. This natural structural design simplifies the filter circuit, is low-cost, cost-effective, stable in performance, and reduces redundancy in the robot’s signal circuit. Integrating this technology with robots can enhance their environmental perception, thereby improving the safety of interactions.

Teacher–student relationships play a vital role in improving college students’ academic performance and the quality of higher education. However, empirical studies with substantial data-driven insights remain limited. To address this gap, this study collected 3278 questionnaires from seven universities across four provinces in China to analyze the key factors affecting college students’ academic performance. A machine learning framework, CQFOA-KELM, was developed by enhancing the Fruit Fly Optimization Algorithm (FOA) with Covariance Matrix Adaptation Evolution Strategy (CMAES) and Quadratic Approximation (QA). CQFOA significantly improved population diversity and was validated on the IEEE CEC2017 benchmark functions. The CQFOA-KELM model achieved an accuracy of 98.15% and a sensitivity of 98.53% in predicting college students’ academic performance. Additionally, it effectively identified the key factors influencing academic performance through the feature selection process.

In the field of medical image processing, combining global and local relationship modeling constitutes an effective strategy for precise segmentation. Prior research has established the validity of Convolutional Neural Networks (CNN) in modeling local relationships. Conversely, Transformers have demonstrated their capability to effectively capture global contextual information. However, when utilized to address CNNs’ limitations in modeling global relationships, Transformers are hindered by substantial computational complexity. To address this issue, we introduce Mamba, a State-Space Model (SSM) that exhibits exceptional proficiency in modeling long-range dependencies in sequential data. Given Mamba’s demonstrated potential in 2D medical image segmentation in previous studies, we have designed a Dual-encoder Global-local Feature Extraction Network based on Mamba, termed DGFE-Mamba, to accurately capture and fuse long-range dependencies and local dependencies within multi-scale features. Compared to Transformer-based methods, the DGFE-Mamba model excels in comprehensive feature modeling and demonstrates significantly improved segmentation accuracy. To validate the effectiveness and practicality of DGFE-Mamba, we conducted tests on the Automatic Cardiac Diagnosis Challenge (ACDC) dataset, the Synapse multi-organ CT abdominal segmentation dataset, and the Colorectal Cancer Clinic (CVC-ClinicDB) dataset. The results showed that DGFE-Mamba achieved Dice coefficients of 92.20, 83.67, and 94.13, respectively. These findings comprehensively validate the effectiveness and practicality of the proposed DGFE-Mamba architecture.

The esophagus is an important part of the human digestive system. Due to its limited regenerative capacity and the infeasibility of donor transplantation, esophageal replacement has become an important problem to be solved urgently in clinics. In recent years, with the rapid development of tissue engineering technology in the biomedical field, tissue engineering stent (artificial esophagus) provides a new therapeutic approach for the repair and reconstruction of esophageal defects and has made remarkable progress. Biomedical esophageal stent materials have also experienced the development process from non-absorbable materials to absorbable materials, and then to new materials with composite cells and biological factors. In this paper, the composition, functional characteristics, and limitations of non-degradable scaffolds, biodegradable scaffolds, and Decellularized Matrix (DM) scaffolds specially designed for these applications are reviewed. Non-absorbable stents are typically composed of synthetic polymers or metals that provide structural support but fail to bind to surrounding tissues over time. In contrast, biodegradable stents are designed to break down gradually in the body while promoting cell infiltration and promoting new tissue formation. DM scaffolds can alleviate autoimmune reactions, preserve natural tissue characteristics, and enable recellularization during auto-repair. In addition, the significance of various cell-loaded materials in esophageal replacement has been explored, and the inclusion of cells in scaffold design has been shown to have the potential to enhance integration with host tissue and improve postoperative functional outcomes. These advances underscore ongoing efforts to closely mimic the structure of the natural esophagus.

This paper proposes the Leg Dimensional Synergistic Optimization Strategy (LDSOS) for humanoid robotic legs based on mechanism decoupling and performance assignment. The proposed method addresses the interdependent effects of dimensional parameters on the local and whole mechanisms in the design of hybrid humanoid robotic legs. It sequentially optimizes the dimensional parameters of the local and whole mechanism, thereby balancing the motion performance requirements of both. Additionally, it considers the assignment of efficient performance resources between the Local Functional Workspace (LFW) and the Whole Available Workspace (WAW). To facilitate the modeling and optimization process, a local/whole Equivalent Configuration Framework (ECF) is introduced. By decoupling the hybrid mechanism into a whole mechanism and multiple local mechanisms, the ECF enhances the efficiency of design, modeling, and performance evaluation. Prototype experiments are conducted to validate the effectiveness of LDSOS. This research provides an effective configuration framework for humanoid robotic leg design, establishing a theoretical and practical foundation for future optimized designs of humanoid robotic legs and pioneering novel approaches to the design of complex hybrid humanoid robotic legs.

In this study, we present a small, integrated jumping-crawling robot capable of intermittent jumping and self-resetting. Compared to robots with a single mode of locomotion, this multi-modal robot exhibits enhanced obstacle-surmounting capabilities. To achieve this, the robot employs a novel combination of a jumping module and a crawling module. The jumping module features improved energy storage capacity and an active clutch. Within the constraints of structural robustness, the jumping module maximizes the explosive power of the linear spring by utilizing the mechanical advantage of a closed-loop mechanism and controls the energy flow of the jumping module through an active clutch mechanism. Furthermore, inspired by the limb movements of tortoises during crawling and self-righting, a single-degree-of-freedom spatial four-bar crawling mechanism was designed to enable crawling, steering, and resetting functions. To demonstrate its practicality, the integrated jumping-crawling robot was tested in a laboratory environment for functions such as jumping, crawling, self-resetting, and steering. Experimental results confirmed the feasibility of the proposed integrated jumping-crawling robot.

The Sine and Wormhole Energy Whale Optimization Algorithm (SWEWOA) represents an advanced solution method for resolving Optimal Power Flow (OPF) problems in power systems equipped with Flexible AC Transmission System (FACTS) devices which include Thyristor-Controlled Series Compensator (TCSC), Thyristor-Controlled Phase Shifter (TCPS), and Static Var Compensator (SVC). SWEWOA expands Whale Optimization Algorithm (WOA) through the integration of sine and wormhole energy features thus improving exploration and exploitation capabilities for efficient convergence in complex non-linear OPF problems. A performance evaluation of SWEWOA takes place on the IEEE-30 bus test system through static and dynamic loading scenarios where it demonstrates better results than five contemporary algorithms: Adaptive Chaotic WOA (ACWOA), WOA, Chaotic WOA (CWOA), Sine Cosine Algorithm Differential Evolution (SCADE), and Hybrid Grey Wolf Optimization (HGWO). The research shows that SWEWOA delivers superior generation cost reduction than other algorithms by reaching a minimum of 0.9% better performance. SWEWOA demonstrates superior power loss performance by achieving ((:{P}_{text{loss,min}})) at the lowest level compared to all other tested algorithms which leads to better system energy efficiency. The dynamic loading performance of SWEWOA leads to a 4.38% reduction in gross costs which proves its capability to handle different operating conditions. The algorithm achieves top performance in Friedman Rank Test (FRT) assessments through multiple performance metrics which verifies its consistent reliability and strong stability during changing power demands. The repeated simulations show that SWEWOA generates mean costs ((:{C}_{text{gen,min}})) and mean power loss values ((:{P}_{text{loss,min}})) with small deviations which indicate its capability to maintain cost-effective solutions in each simulation run. SWEWOA demonstrates great potential as an advanced optimization solution for power system operations through the results presented in this study.

Meta-heuristic evolutionary algorithms have become widely used for solving complex optimization problems. However, their effectiveness in real-world applications is often limited by the need for many evaluations, which can be both costly and time-consuming. This is especially true for large-scale transportation networks, where the size of the problem and the high computational cost can hinder the algorithm’s performance. To address these challenges, recent research has focused on using surrogate-assisted models. These models aim to reduce the number of expensive evaluations and improve the efficiency of solving time-consuming optimization problems. This paper presents a new two-layer Surrogate-Assisted Fish Migration Optimization (SA-FMO) algorithm designed to tackle high-dimensional and computationally heavy problems. The global surrogate model offers a good approximation of the entire problem space, while the local surrogate model focuses on refining the solution near the current best option, improving local optimization. To test the effectiveness of the SA-FMO algorithm, we first conduct experiments using six benchmark functions in a 50-dimensional space. We then apply the algorithm to optimize urban rail transit routes, focusing on the Train Routing Optimization problem. This aims to improve operational efficiency and vehicle turnover in situations with uneven passenger flow during transit disruptions. The results show that SA-FMO can effectively improve optimization outcomes in complex transportation scenarios.