In the task of automatic sleep stage classification, deep learning models often face the challenge of balancing temporal-spatial feature extraction with computational complexity. To address this issue, this study introduces FlexibleSleepNet, a lightweight convolutional neural network architecture designed around the Adaptive Feature Extraction (AFE) Module and Scale-Varying Compression (SVC) Module. Through multi-channel polysomnography data input and preprocessing, FlexibleSleepNet utilizes the AFE Module to capture intra-channel features and employs the SVC Module for channel feature compression and dimension expansion. The collaborative work of these modules enables the network to effectively capture temporal-spatial dependencies between channels. Additionally, the network extracts feature maps through four distinct stages, each from different receptive field scales, culminating in precise sleep stage classification via a classification module. This study conducted k- fold cross-validation on three different databases: SleepEDF-20, SleepEDF-78, and SHHS. Experimental results show that FlexibleSleepNet demonstrates superior classification performance, achieving classification accuracies of 86.9% and 87.6% on the SleepEDF-20 and SHHS datasets, respectively. It performs particularly well on the SleepEDF-78 dataset, where it reaches a classification accuracy of 87.0%. Additionally, it has significantly enhanced computational efficiency while maintaining low computational complexity.

Bimanual telemanipulation is vital for facilitating robots to complete complex and dexterous tasks that involve two handheld objects under teleoperation scenarios. However, the bimanual configuration introduces higher complexity, dynamics, and uncertainty, especially in those uncontrolled and unstructured environments, which require more advanced system integration. This paper presents a bimanual robotic teleoperation architecture with modular anthropomorphic hybrid grippers for the purpose of improving the telemanipulation capability under unstructured environments. Generally, there are two teleoperated subsystems within this architecture. The first one is the Leap Motion Controller and the anthropomorphic hybrid robotic grippers. Two 3D printed anthropomorphic hybrid robotic grippers with modular joints and soft layer augmentations are designed, fabricated, and equipped for telemanipulation tasks. A Leap Motion Controller is used to track the motion of two human hands, while each hand is utilized to teleoperate one robotic gripper. The second one is the haptic devices and the robotic arms. Two haptic devices are adopted as the master devices while each of them takes responsibility for one arm control. Based on such a framework, an average RMSE (root-mean-square-error) value of 0.0204 rad is obtained in joint tracking. Nine sign-language demonstrations and twelve object grasping tasks were conducted with the robotic gripper teleoperation. A challenging bimanual manipulation task for an object with 5.2 kg was well addressed using the integrated teleoperation system. Experimental results show that the proposed bimanual teleoperation system can effectively handle typical manipulation tasks, with excellent adaptabilities for a wide range of shapes, sizes, and weights, as well as grasping modes.

Intestinal-related diseases all around the world are increasing nowadays, and gradually become stubborn diseases threatening human health, and even lives. Diagnosis methods have attracted more and more attention. This article concerns a non-invasive way, a novel micro-robot, to diagnose intestinal diseases. This proposed micro-robot is a swallowable device, 14 mm in diameter, like a capsule. In order to make it possible for the micro-robot to move forward, backward, or anchor itself at a suspicious lesion point in the intestine with different lumen diameter sections, two key mechanisms have been proposed. One is an expanding mechanism with two-layer folded legs for anchoring. The designed expanding mechanism could realize a large variable diameter ratio, upwards of 3.43. In addition, a pair of specific annular gears instead of a traditional pinion drive is devised not only saving limited space, but also reducing energy loss. The other mechanism is a telescoping mechanism, possessing a self-locking lead screw nut system, which is used to obtain axial motion of the micro-robot. Then, the kinematics and dynamics of the micro-robot are analyzed. After that, the following experiments, including force tests and locomotion tests, are constructed. A good match is found between the theoretical results and the experimental data. Finally, in vitro experiments are performed with a prototype to verify the safety and reliability of the proposed micro-robot in porcine intestine.

Topology optimization (TO) is a powerful generative design tool for innovative structural design, capable of optimizing material distribution to generate structures with superior performance. However, current topology optimization algorithms mostly target a single objective and are highly dependent on the problem definition parameters, causing two critical issues: limited human controllability and solution diversity. These issues often lead to burdensome design iterations and insufficient design exploration. This paper proposes a multi-solution TO framework to address them. Human designers express their stylistic preferences for structures through sketches which are decomposed into stroke and closed-shape elements to flexibly guide each TO process. Sketch-based constraints are integrated with Fourier mapping-based length-scale control to enhance human controllability. Solution diversity is achieved by perturbing Fourier mapping frequencies and load conditions in the neural implicit TO framework. Adaptive parallel scale adjustment is incorporated to reduce the computational cost for design exploration. Using the structural design of a wheel spoke as a case study, the mechanical performance and diversity of the generated TO solutions as well as the effectiveness of human control are analyzed both qualitatively and quantitatively. The results reveal that the sketch-based constraints and length-scale control have distinct control effects on structural features and have different impacts on the mechanical performance and diversity, thereby enabling fine-grained and flexible human controllability to better balance conflicting objectives.

This paper introduces the Wavelet-Enhanced Swin Transformer Network (WSC-Net), a novel dual-branch architecture that resolves the inherent tradeoff between global spatial contextual and fine-grained spectral details in hyperspectral image (HSI) classification. While transformer-based models excel at capturing long-range dependencies, their patch-based nature often overlooks intra-patch high-frequency details, hindering the discrimination of spectrally similar classes. Our framework synergistically couples a two-stage Swin transformer with a parallel Wavelet Transform Module (WTM) for local frequency information capture. To address the semantic gap between spatial and frequency domains, we propose the Cross-Domain Attention Fusion (CDAF) module—a bi-directional attention mechanism that facilitates intelligent feature exchange between the two streams. CDAF explicitly models cross-domain dependencies, amplifies complementary features, and suppresses noise through attention-guided integration. Extensive experiments on four benchmark datasets demonstrate that WSC-Net consistently outperforms state-of-the-art methods, confirming its effectiveness in balancing global contextual modeling with local detail preservation.