AI is increasingly being used in stroke rehabilitation to predict recovery trajectories and monitor motor function, providing adaptive and continuous care. The review highlights the need for standardized outcome measures and data pipelines to determine the clinical impact of AI-based monitoring.
A meter-scale flapping-wing robot demonstrates the benefits of low-angle takeoff, increasing flight distance by 49% and horizontal distance by 123%. The robot's ability to control pitch attitude through the flapping-pitch effect enables early acceleration and speed advantage.
Researchers developed a new AI search framework, CMS-NAS, to predict epileptic seizures with improved accuracy and efficiency. The framework balances prediction performance and model size, achieving state-of-the-art results on public datasets.
A new AI-based multimorbidity score, LLM-MMS, accurately predicts mortality risk for individuals with multimorbidity, outperforming traditional comorbidity indices in transportability and calibration. The model's use of large language models integrates diverse health information, providing a holistic phenotype representation.
A study on on-orbit autonomous assembly of deployable modules reveals the importance of real-time deformation control and motion control during assembly. Numerical simulations demonstrate the effectiveness of a proposed control strategy, reducing geometric errors to below 0.013 mm.
A bioinspired perceptual sensor was developed to decode haptic and pain stimuli in prostheses, integrating a piezoelectric haptic pathway with a neuromorphic pain pathway. The system provides real-time haptic localization and spatiotemporal pain accumulation, supporting closed-loop robotic avoidance and prosthetic rehabilitation feedback.
Researchers developed a method to extract scintillation amplitude errors from amplitude streaks and retrieve ionospheric parameters using L-band SAR. The study reveals high-resolution observations of ionospheric scintillation and its relationship with ionospheric anomalies.
A closed-loop modeling and assessment framework for aerospace engineering managers' AI competency is proposed, capturing governance priorities and multi-attribute interdependencies. The framework, comprising five dimensions and twenty attributes, provides a traceable decision-support tool for selection, performance management, and care...
Researchers propose a bi-objective optimization framework for VLEO satellites based on a data-relay multi-satellite system, balancing computational feasibility and control precision. The proposed method outperforms conventional controls in terms of attitude control accuracy and extends the downlink window through relay satellites.
Researchers propose an error-state extended Kalman filtering method for state estimation in flexible cable-net recovery systems. The method directly employs multibody integrators for one-step state prediction, improving stability and accuracy, and achieving real-time operational capability.
Researchers propose a transfer path design method from cislunar orbits to retrograde geosynchronous orbits, revealing four transfer types and optimizing fuel consumption. Simulation results show that three-impulse solutions can achieve low-energy transfers within 8 days, with minimum fuel consumption below 1.5 km/s.
The LUMIO CubeSat will observe Apophis's 2029 flyby from the Earth-moon L2 point, providing valuable data on tidal interactions and asteroid observation. The study shows that LUMIO-Cam can detect Apophis as a sub-pixel target and observe its spin-period information.
The proposed framework integrates knowledge graphs and large language models to diagnose space TWTAs, achieving high adoption rates and robust stability, with a 94.56% adoption rate and 95% accuracy in real-scenario testing.
Carbon-based materials exhibit ultra-high thermal conductivity, low density, and low coefficient of thermal expansion, making them suitable for space optical remote sensors. Graphite materials show significant promise in heat dissipation and lightweighting, while carbon/carbon composites improve temperature homogenization.
A new hierarchical path planning method optimizes internal and external parameters of Dynamic Movement Primitives to ensure precise assembly and safe obstacle avoidance. The method achieves minimum obstacle clearance of 0.078m and 0.018m in static and dynamic scenarios, respectively.
Researchers develop attitude-controllable radar image generation method using ControlNet, achieving better SSIM, PSNR, and FID scores than existing methods. The method reduces image entropy and generates clearly contoured images with precise control over elevation angle.
Intracortical microstimulation (ICMS) evokes perception and plasticity in brain-computer interfaces, creating structured sensory information and potentially guiding behavior. The technology induces plasticity-like changes in cortical networks, with implications for neural repair and repair.
Researchers at Beijing Jiaotong University designed an electric ignition system for an ammonium dinitramide-based thruster, eliminating the need for catalysts and preheating. The system achieved cold-start ignition and stable combustion at room temperature, with optimal ignition voltage identified as 80 V.
A new hierarchical motion planning method for space redundant manipulators is proposed, integrating goal-biased RRT with sequential convex programming. The method generates smooth trajectories satisfying obstacle avoidance, end-effector line-of-sight constraints, and joint torque limits, converging in 17 and 18 iterations for static an...
The IncResUnet model detects plasma bubbles in the ionosphere with high accuracy, outperforming traditional threshold methods. It achieves an F1-score of 0.914 and recall of 0.958 on the FORMOSAT-1 test set, making it an efficient automated detection tool for space weather monitoring and ionospheric physics research.
A study using machine learning and multi-omics profiling found a consistent association between lower bone mineral density and higher age-related macular degeneration risk. Shared molecular signatures and pathways involving extracellular matrix remodeling, lipid transport, and inflammation were also identified.
Researchers propose a pre-rerouting strategy to minimize service disruption and resource overhead in LEO satellite networks. The strategy leverages an extended OSPF protocol to synchronize link state databases and redefine routing metrics, enabling seamless service switching before predictable link interruptions.
A new numerical method for binary asteroid systems uses Hamel's variational integrators to preserve symplectic structure and Lie group geometric properties. Simulation results show improved energy conservation and orthogonality preservation compared to existing methods.
A study proposes a multi-model combined observation method for non-cooperative spacecraft using sequential coalition game theory. The method generates an optimal observation strategy sequence under minimum fuel constraints, improving space security assessment capabilities.
Researchers propose a novel navigation method for high-altitude, high-speed vehicles using relative position vector measurements. The method achieves three-dimensional positioning accuracy better than 200m over 150 minutes, outperforming existing navigation methods.
Researchers develop a predictive learning-based pursuit strategy for the multiple-to-one orbital pursuit-evasion game, achieving a 95% capture success rate. The strategy combines evader action prediction with the Deep Deterministic Policy Gradient algorithm, converging faster and achieving higher rewards than conventional methods.
Researchers propose an adaptive dueling double deep Q-network (D3QN) for active defense guidance, addressing partial observability and multi-strategy adversarial conditions. The approach achieves high evasion success rates and robustness under various conditions.
Research proposes a developmental roadmap for lunar robots, categorizing them into programmed and intelligent eras, with a focus on 'intelligence + new energy' empowerment. A comprehensive technological framework is established, encompassing key technologies such as environmental perception, autonomous decision-making, and precision ma...
This study proposes BioPINN-LM, a hybrid framework for rapid wall-stress prediction and conversational risk interpretation in ascending thoracic aortic aneurysms. The framework outperforms baseline models and achieves high accuracy in wall-stress prediction and conversational decision support.
Researchers investigated the influence of thermocapillary convection on phase-change material melting under varying gravity conditions. The study found that aspect ratio is a key factor in determining the dominant convective regime, with thermocapillary effects dominating in flat cavities and natural convection dominating in cavities w...
The study proposes an electric ignition scheme for ADN-based thrusters, eliminating the need for catalysts and preheating, and achieving rapid cold-start ignition. The results show that the thruster achieves stable combustion at room temperature, with optimal ignition voltage identified as 80 V.
A deep learning model called IncResUnet detects plasma bubbles in the ionosphere with high accuracy. The model achieved an F1-score of 0.914 on the FORMOSAT-1 test set and outperformed traditional threshold methods.
Researchers propose a new numerical method for binary asteroid systems, achieving numerical accuracy and structure preservation. The method outperforms existing methods in energy conservation and orthogonality preservation, making it suitable for long-term predictions and planetary defense mission planning.
A new pre-rerouting strategy ensures seamless service continuity in LEO satellite networks by redefining routing metrics to account for link effective duration and available bandwidth. The strategy achieves a packet loss rate of 1% compared to 7% with conventional rerouting schemes.
The Apophis asteroid will make a rare 38,000 km flyby of Earth in 2029, offering a unique chance for planetary science and defense. The flyby will provide direct evidence for the structural evolution of rubble-pile asteroids and demonstrate the feasibility of various exploration modes.
A novel hierarchical motion planning method combines goal-biased RRT with sequential convex programming to plan high-quality trajectories for space redundant manipulators. The proposed method successfully satisfies obstacle avoidance, end-effector line-of-sight constraints, and joint torque limits, converging in 17 and 18 iterations fo...
A new strategy design method for multi-aspect information acquisition of non-cooperative targets is proposed using sequential coalition game theory. The method combines four close-range operation models to achieve efficient acquisition of multi-aspect payload information under minimum fuel constraints.
The 2029 Apophis flyby will offer a unique opportunity for planetary science and defense against a potentially hazardous asteroid. The event will provide insights into Apophis' physical state and expected effects on its orbit and spin, as well as potential surface material displacement and internal structural response.
The bionic multichannel whisker system provides spatial coverage for fine surface interrogation and radial collision detection, improving shape reconstruction and force estimation. The system's calibration framework and low-computational-overhead correction ensure cross-channel consistency and reliability.
A single-source dual-drive flexible knee assistive exoskeleton was designed to support older adults in real daily walking scenarios. The system can recognize multiple daily locomotion patterns and provide appropriate assistive torque in real time, reducing users' movement burden and improving balance and mobility.
A wearable ultrasound patch with a semi-supervised segmentation network and a dedicated clinical predictor is developed for noninvasive central venous pressure monitoring in ICU patients. The system achieved high accuracy in detecting elevated CVP, with IJV Max Area and IJV Ratio as dominant predictors.
The robot combines two biomimetic strategies to achieve exceptional speed, load capacity, and maneuverability. It maintains high speed under heavy loads and has impressive energy efficiency.
A new AI-driven EMG control framework has been developed to improve elderly assistance. The framework uses a 3-level approach to translate EMG signals into robotic assistive actions across various levels. It shows improved results in physiological signal labeling, functional multijoint intent decoding, and behavioral task adaptation.
A recent study found that recurrent heatwave exposure is linked to accelerated biological aging in middle-aged and older adults. The research, using data from the China Health and Retirement Longitudinal Study, discovered that increased heatwave frequency and duration lead to disruptions in lipid and glucose metabolism, ultimately acce...
During early simulated microgravity, cerebral blood flow changes are closely associated with upstream macrovascular hemodynamics in internal carotid and vertebral arteries. The study found that regional cerebral blood flow alterations are independently predicted by vertebral artery hemodynamic parameters.
Researchers propose a multiple-moon-aided Jovian capture method utilizing solar gravity perturbation (SGP) to reduce velocity increments. The study uses four dynamic models and computes datasets to analyze the mechanism of SGP, revealing its influence on perijove radius change.
A study proposes two detonation defense modes to defend against large-sized near-Earth asteroids. The novel flyby pre-excavation detonation mode offers stronger energy coupling and can destroy hundred-meter-scale asteroids, making it the preferred option when warning time permits.
The review explores various propulsion strategies, navigation algorithms, and on-demand release mechanisms for autonomous microrobots. The authors propose a four-tier hierarchy based on barrier traversal capability, enabling local, systemic, deep tissue, and immune-privileged delivery.
Journal Cyborg and Bionic Systems achieved a ranking of 2nd in Robotics and 4th in Engineering, Biomedical in the Journal Citation Report 2025. The journal has an Impact Factor of 20.9 and is published by Beijing Institute of Technology Press Co., Ltd.
Optoelectronic tweezers (OETs) manipulate single cells through dielectrophoretic forces, offering a unique approach to cell capture, transport, sorting, analysis, and functional screening. OET integrates with microfluidics and other technologies for label-free single-cell physical parameter measurement and biomolecule manipulation.