Journal of Bionic Engineering最新文献

As an effective therapy for treating unilateral neglect, Mirror Therapy (MT) is employed in the upper limb motor function rehabilitation of hemiplegic patients. However, traditional MT has a serious limitation—the Impaired Limb (IL) doesn’t actually move. In this study, a novel performance-based assistance strategy suitable for Robotic Mirror Therapy (RMT) based on MT is proposed. A guiding assistance based on the progress difference HL and IL is constructed in trajectory guidance, and a multi-stiffness region correction force field based on trajectory tracking error is designed to constrain IL’s deviation from the intended path in trajectory correction assistance. To validate the presented strategy, a series of experiments on a RMT system based on the end-effector upper limb rehabilitation robot are conducted. The results verify the performance and feasibility of this strategy.

Runge Kutta Optimization (RUN) is a widely utilized metaheuristic algorithm. However, it suffers from these issues: the imbalance between exploration and exploitation and the tendency to fall into local optima when it solves real-world optimization problems. To address these challenges, this study aims to endow each individual in the population with a certain level of intelligence, allowing them to make autonomous decisions about their next optimization behavior. By incorporating Reinforcement Learning (RL) and the Composite Mutation Strategy (CMS), each individual in the population goes through additional self-improvement steps after completing the original algorithmic phases, referred to as RLRUN. That is, each individual in the RUN population is trained intelligently using RL to independently choose three different differentiation strategies in CMS when solving different problems. To validate the competitiveness of RLRUN, comprehensive empirical tests were conducted using the IEEE CEC 2017 benchmark suite. Extensive comparative experiments with 13 conventional algorithms and 10 advanced algorithms were conducted. The experimental results demonstrated that RLRUN excels in convergence accuracy and speed, surpassing even some champion algorithms. Additionally, this study introduced a binary version of RLRUN, named bRLRUN, which was employed for the feature selection problem. Across 24 high-dimensional datasets encompassing UCI datasets and SBCB machine learning library microarray datasets, bRLRUN occupies the top position in classification accuracy and the number of selected feature subsets compared to some algorithms. In conclusion, the proposed algorithm demonstrated that it exhibits a strong competitive advantage in high-dimensional feature selection for complex datasets.

This paper investigates the motion control of the heavy-duty Bionic Caterpillar-like Robot (BCR) for the maintenance of the China Fusion Engineering Test Reactor (CFETR). Initially, a comprehensive nonlinear mathematical model for the BCR system is formulated using a physics-based approach. The nonlinear components of the model are compensated through nonlinear feedback linearization. Subsequently, a fuzzy-based regulator is employed to enhance the receding horizon optimization process for achieving optimal results. A Deep Neural Network (DNN) is trained to address disturbances. Consequently, a novel hybrid controller incorporating Nonlinear Model Predictive Control (NMPC), the Fuzzy Regulator (FR), and Deep Neural Network Feedforward (DNNF), named NMPC-FRDNNF is developed. Finally, the efficacy of the control system is validated through simulations and experiments. The results indicate that the Root Mean Square Error (RMSE) of the controller with FR and DNNF decreases by 33.2 and 48.9%, respectively, compared to the controller without these enhancements. This research provides a theoretical foundation and practical insights for ensuring the future highly stable, safe, and efficient maintenance of blankets.

Soft grippers are favored for handling delicate objects due to their compliance but often have lower load capacities compared to rigid ones. Variable Stiffness Module (VSM) offer a solution, balancing flexibility and load capacity, for which particle jamming is an effective technology for stiffness-tunable robots requiring safe interaction and load capacity. Specific applications, such as rescue scenarios, require quantitative analysis to optimize VSM design parameters, which previous analytical models cannot effectively handle. To address this, a Grey-box model is proposed to analyze the mechanical response of the particle-jamming-based VSM by combining a White-box approach based on the virtual work principle with a Black-box approach that uses a shallow neural network method. The Grey-box model demonstrates a high level of accuracy in predicting the VSM force-height mechanical response curves, with errors below 15% in almost 90% of the cases and a maximum error of less than 25%. The model is used to optimize VSM design parameters, particularly those unexplored combinations. Our results from the load capacity and force distribution comparison tests indicate that the VSM, optimized through our methods, quantitatively meets the practical engineering requirements.

Passively stabilized double-wing Flapping Micro Air Vehicles (FMAVs) do not require active control and exhibit good electromagnetic interference resistance, with significant research value. In this paper, the dynamic model of FMAV was established as the foundation for identifying flapping damping coefficients. Through a pendulum experiment, we ascertain the flapping damping of the damper using the energy conservation method. Besides, fitting relationships between the damper area, damper mass, and the moment of inertia are developed. The factors influencing the bottom damper damping are determined using correlation coefficients and hypothesis testing methods. Additionally, stable dampers are installed on both the top and bottom of the FMAV to achieve passive stability in simulations. The minimum damper areas for the FMAV were optimized using genetic algorithms, resulting in a minimum top damper area of 128 cm(^{2}) and a minimum bottom damper area of 80 cm(^{2}). A prototype with a mass of 25.5 g and a wingspan of 22 cm has been constructed. Prototype testing demonstrated that FMAV can take off stably with a 3 g payload and a tilt angle of 5(^{circ }). During testing, the area-to-mass ratio of the FMAV reached 7.29 cm(^{2})/g, achieving passive stability with the world’s smallest area-to-mass ratio.

The second-leading cause of cancer-related deaths globally is liver cancer. The treatment of liver cancers depends heavily on the accurate segmentation of liver tumors from CT scans. The improved method based on U-Net has achieved good performance for liver tumor segmentation, but these methods can still be improved. To deal with the problems of poor performance from the original U-Net framework in the segmentation of small-sized liver tumors and the position information of tumors that is seriously lost in the down-sampling process, we propose the Multi-attention Perception-fusion U-Net (MAPFUNet). We propose the Position ResBlock (PResBlock) in the encoder stage to promote the feature extraction capability of MAPFUNet while retaining the position information regarding liver tumors. A Dual-branch Attention Module (DWAM) is proposed in the skip connections, which narrows the semantic gap between the encoder's and decoder's features and enables the network to utilize the encoder's multi-stage and multi-scale features. We propose the Channel-wise ASPP with Attention (CAA) module at the bottleneck, which can be combined with multi-scale features and contributes to the recovery of micro-tumor feature information. Finally, we evaluated MAPFUNet on the LITS2017 dataset and the 3DIRCADB-01 dataset, with Dice values of 85.81 and 83.84% for liver tumor segmentation, which were 2.89 and 7.89% higher than the baseline model, respectively. The experiment results show that MAPFUNet is superior to other networks with better tumor feature representation and higher accuracy of liver tumor segmentation. We also extended MAPFUNet to brain tumor segmentation on the BraTS2019 dataset. The results indicate that MAPFUNet performs well on the brain tumor segmentation task, and its Dice values on the three tumor regions are 83.27% (WT), 84.77% (TC), and 76.98% (ET), respectively.

Microscopy, crucial for exploring biological structures, often uses polarizing microscopes to observe tissue anisotropy and reconstruct label-free images. However, these images typically show low contrast, and while fluorescence imaging offers higher contrast, it is phototoxic and can disrupt natural assembly dynamics. This study focuses on reconstructing fluorescence images from label-free ones using a complex nonlinear transformation, specifically aiming to identify organelles within diverse optical properties of tissues. A multimodal deep learning model, 3DTransMDL, was developed, employing the Stokes vector to analyze the sample’s retardance, phase, and orientation. This model incorporates isotropy and anisotropy to differentiate organelles, enhancing the input with structures' varied optical properties. Additionally, techniques like background distortion normalization and covariate shift methods were applied to reduce noise and overfitting, improving model generalization. The approach was tested on mouse kidney and human brain tissues, successfully identifying specific organelles and demonstrating superior performance in reconstructing 3D images, significantly reducing artifacts compared to 2D predictions. Evaluation metrics such as SSIM, PCC, and R2 score confirm the model's efficacy, with improvements observed in multi-modality input setups. This advancement suggests potential applications in molecular dynamics, aiming for further enhancements in future studies.

The welding of medium and thick plates has a wide range of applications in the engineering field. Industrial welding robots are gradually replacing traditional welding operations due to their significant advantages, such as high welding quality, high work efficiency, and effective reduction of labor intensity. Ensuring the accuracy of the welding trajectory for the welding robot is crucial for guaranteeing welding quality. In this paper, the author uses the chaos sparrow search algorithm to optimize the trajectory of a multi-layer and multi-pass welding robot for medium and thick plates. Firstly, the Sparrow Search Algorithm (SSA) is improved by introducing tent chaotic mapping and Gaussian mutation of the inertia weight factor. Secondly, in order to prevent the welding robot arm from colliding with obstacles in the welding environment during the welding process, maintain the stability of the welding robot, and ensure the continuous stability of the changes in each joint angle, joint angular velocity, and angular velocity of the joint angle, a welding robot model is established by improving the Denavit–Hartenberg parameter method. A multi-objective optimization fitness function is used to optimize the trajectory of the welding robot, minimizing time and energy consumption. Thirdly, the optimization and convergence performance of SSA and Chaos Sparrow Search Algorithm (CSSA) are compared through 10 benchmark test functions. Based on the six sets of test functions, the CSSA algorithm consistently maintains superior optimization performance and has excellent stability, with a faster decline in the convergence curve compared to the SSA algorithm. Finally, the accuracy of welding is tested through V-shaped multi-layer and multi-pass welding experiments. The experimental results show that the CSSA algorithm has a strong superiority in trajectory optimization of multi-layer and multi-pass welding for medium and thick plates, with an accuracy rate of 99.5%. It is an effective optimization method that can meet the actual needs of production.

For promising applications such as soft robotics, flexible haptic monitors, and active biomedical devices, it is important to develop ultralow voltage, highly-performant artificial muscles with high bending strains, rapid response times, and superior actuation endurance. We report a novel highly performant and low-cost artificial muscle based on microfibrillated cellulose (MFC), ionic liquid (IL), and polyvinyl alcohol (PVA), The proposed MFC–IL–PVA actuator exhibits excellent electrochemical performance and actuations characteristics with a high specific capacitance of 225 mF/cm2, a large bending strain of 0.51%, peak displacement up to 7.02 mm at 0.25 V ultra-low voltage, outstanding actuation flexural endurance (99.1% holding rate for 3 h), and a wide frequency band (0.1–5 Hz). These attributes stem mainly from its high specific surface area and porosity, tunable mechanical properties, and the strong ionic interactions of cations and anions with MFC and PVA in ionic liquids. Furthermore, bionic applications such as bionic flytraps, bionic butterflies with vibrating wings, and smart circuit switches have been successfully realized using this technology. These specific bionic applications demonstrate the versatility and potential of the MFC–IL–PVA actuator, highlighting its important role in the fields of bionic engineering, robotics, and smart materials. They open up new possibilities for innovative scientific research and technological applications.

4D printed smart materials is mostly relying on thermal stimulation to actuate, limiting their widely application requiring precise and localized control of the deformations. Most existing strategies for achieving localized control rely on heterogeneous material systems and structural design, thereby increasing design and manufacturing complexity. Here, we endow localized electrothermal, actuation, and sensing properties in electrically-driven soft actuator through parameter-encoded 4D printing. We analyzed the effects of printing parameters on shape memory properties and conductivity, and then explored the multi-directional sensing performance of the 4D printed composites. We demonstrated an integrated actuator-sensor device capable of both shape recovery and perceiving its own position and obstacles simultaneously. Moreover, it can adjust its sensing characteristics through temporary shape programming to adapt to different application scenarios. This study achieves integrated and localized actuation-sensing without the need for multi-material systems and intricate structural designs, offering an efficient solution for the intelligent and lightweight design in the fields of soft robotics, biomedical applications, and aerospace.