Researchers developed V-UNet, a novel model combining global and local information to address noise and redundant information in medical images. The model achieved competitive segmentation performance while maintaining low computational requirements, promising a more efficient and robust AI-assisted medical image analysis.
A new framework jointly optimizes UAV trajectories and FANET topology to maximize data transmission, outperforming existing methods in field experiments and simulations. The approach enables more efficient and reliable multi-UAV missions for environmental monitoring and other applications.
Researchers develop a novel framework, LL-Refiner, to enhance high-resolution images in poor lighting conditions, outperforming state-of-the-art techniques. The framework uses a coarse enhancement stage to guide the recovery of fine details, resulting in improved visual quality and performance in downstream computer-vision tasks.
Researchers present a novel protocol structure for achieving global/semi-global finite-time consensus in multi-agent systems. The protocols use a hyperbolic tangent function to guarantee consensus and provide explicit calculation of settling time, making them practical for real-world applications.
A novel technique has been proposed to enhance the learning ability of robots performing repetitive tasks by using a fractional power update rule. The study demonstrates fast convergence rates and potential applications in industries such as autonomous vehicles and rehabilitation robots.
A new CNN framework, PE-Net, is proposed for predicting machine remaining useful life (RUL) accurately. The framework uses a novel architecture with small-sized one-dimensional convolution kernels and deep networks to learn features from input time series signals.
A recent review article summarizes the latest developments in finite-control-set model predictive control (FCS-MPC) strategies for PMSMs. FCS-MPC is a promising approach to optimize drive systems, but challenges remain, including computational complexity and parameter uncertainty.
Researchers used game theory to create models of cooperative and competitive behaviors in multi-agent systems, focusing on distributed online optimization, federated optimization, and static/dynamic games. The findings have potential applications in smart cities, market competition, information security, and drug development.
Researchers at Michigan State University have designed a learning-enabled safe controller for systems operating in uncertain environments. The new method, which combines control barrier functions and Lyapunov functions, allows the system to quickly learn uncertainties while achieving maximum safe performance.
A new control allocation method using a neural network improves the performance of quadrotor controllers by considering aerodynamic effects. This approach reduces errors in command generation and delivers better thrust and torque signals.
A new approach uses reinforcement learning algorithm to help robotic knee mimic intact human knee in walking, achieving 100% success rate on even ground. The technology also adapts to uneven terrain and changes in walking pace, promising a more comfortable experience for prosthetic users.
Researchers developed a machine learning approach enabling robots to separate, recognize, and grasp individual objects with high accuracy. The method achieved 97% success rate in real-world experiments, paving the way for industrial parts sorting and residential waste sorting applications.
A novel dynamic event-triggered scheduling approach is proposed to solve the platooning control problem, demonstrating effective trade-off between performance and efficient communication. The researchers aim to further investigate resource-efficient control strategies to preserve satisfactory operational performance.
Researchers have developed a novel algorithm that improves the speed at which nodes converge on agreement regarding a single data value needed during computation. The technique is based on designing an optimized network topology that reduces communication costs while allowing for adjustable convergence rates.
The robotic white cane system combines depth data with a 2D floor plan map to reduce pose estimation errors. It features a novel 'robotic roller tip' interface that allows for automatic mode-switching, making it easier for visually impaired users to navigate.
A team of researchers has created a new algorithm that determines the most efficient route for robots to navigate complex spaces. The RBF-Galerkin method combines two existing approaches to find the optimal solution, surpassing other methods in terms of cost and time efficiency.
Drones are not the only vehicles at the forefront of automating deliveries; researchers have developed a control strategy for ships and partnered tugs to navigate environmental disturbances. The multi-layer, multi-agent control scheme increases efficiency in the shipping process while reducing human error.
AI is transforming various aspects of life, such as road safety with EEG-based driver state estimation techniques and automatic image captioning for indexing large datasets. It also enhances surveillance tasks with cognitive memory-augmented networks, improving accuracy and reliability.
A robot system with a digital twin uses intra-operative ultrasound to guide cardiac surgeons, improving their view of the patient without radiation exposure. The platform can assist less experienced operators and aid in pre-planning and real-time control.
Researchers developed a new modeling system to enhance dynamic positioning, allowing ships to stay fixed in challenging sea conditions. The approach includes a digital observer that translates wind or wave disturbances into specific measurements, enabling the vessel to respond in a reasonable timeframe.
A team of researchers developed a method to detect simultaneous sensor and actuator faults in wind turbines, eliminating the need for redundant hardware components. The approach uses a state observer model to identify discrepancies between the original system and its duplicate, enabling real-time fault detection and correction.
A research team from Iowa State University has developed a way to control hard-to-predict systems. The technique uses quotienting to identify the least fixed-point operator, which can result in a new model that acts as a supervisor of the system.
Researchers are studying the integration of smart vehicles with smart cities, exploring computing paradigms such as vehicular cloud computing and vehicular fog computing. The goal is to create efficient, safe, and economical transportation systems that can adapt to real-time applications.
A new approach to solving multiagent decision making problems using agent-by-agent optimization leads to reduced computational complexity, performing at par with standard rollout algorithms. The result is a dramatic reduction in computation cost with linear growth in computation with the number of agents.
Researchers from the University of Calabria developed a predictive control scheme that can identify and protect against replay attacks in distributed networks. The approach uses a 'receding horizon' model to predict future system behavior and detect unexpected events, allowing for swift protection against malicious actors.
Researchers developed a digital twin environment that mirrors physical welding setups to train new users and protect physical systems. The system tracks welders' behavior patterns, enabling efficient novice training and safe practice without risk of damage.
Researchers developed an AI model to detect COVID-19 in chest X-rays with high accuracy, using a large dataset of other X-ray images as a starting point. The tool has the potential to assist doctors in identifying, measuring the severity and classifying the disease.
Researchers in Italy and UK developed an automatic process to assess nanofiber fabrication quality, achieving 92.5% accuracy. The new system reduces the need for human inspection and minimizes anomalies, improving uniformity and quality.
Researchers have developed an automated silicon-substrate ultra-microtome to improve the speed and quality of brain neural connection reconstruction. The device reduces manual collection skill requirement and ensures high-quality imaging without post-processing operations.
Researchers propose a novel architecture, Med-BDA, to analyze healthcare big data, enabling real-time predictions and better patient treatments. The new approach uses Apache Spark technology to tackle complex data analysis challenges.
A comprehensive analysis identifies risks in smart agriculture and proposes countermeasures to prevent information theft and cyberattacks. The researchers categorized agricultural systems into three modes and proposed six general countermeasures, including technological and physical solutions.
A proposed algorithm could reduce carbon dioxide emissions from semi-trucks and other heavy-duty vehicles by up to 9.3%. The Cloud Computing System uses data from external sources like road slope, speed limits, and weather conditions to determine the most fuel-efficient route.
Researchers from India and the UK developed a new brain-computer interface that uses EEG signals to control a robot arm with reduced positional error. The system achieves this by utilizing the P300 signal, which allows the robotic arm to make finer adjustments and reduce errors.
Researchers have designed a cloud-based autonomous system framework utilizing the standard messaging protocol for IoT. This framework maximizes network coverage area and increases speed of communication between unmanned sensors. The team plans to add computer vision and machine learning capabilities in future developments.
Researchers developed a novel control architecture that defends complex, interconnected systems against cyberattacks by implementing a Leader-Follower approach. This algorithm can detect and isolate infected sub-systems, reducing the impact of targeted DoS attacks and increasing system robustness.
Researchers have developed an efficient method to estimate camera movement, reducing the number of hypotheses generated from up to five to one. The new approach allows for real-time execution of pose estimation, with a complete algorithm taking only 29 milliseconds per frame.
A team of researchers from Italy has proposed a new metric to evaluate the effectiveness of virtual groups. By analyzing interaction data from two Italian-based social networks, they found that trust plays a crucial role in forming cohesive and productive groups.
Researchers developed a single-camera machine vision algorithm allowing indoor robots to guide themselves by identifying reference points on a tiled floor. The technology has wide-ranging potential applications in warehouses, distribution centers, and industrial settings.
Researchers have developed a new feedforward method that improves on conventional techniques by obtaining parameters from an uncertain environment. This approach achieved better performance than the traditional method in simulations, and is set to be tried out on industrial robots and machine vision.
Researchers developed a novel AI-managed trading strategy that outperforms traditional methods, achieving greater gains and fewer losses. The proposed system utilizes convolutional neural networks to analyze layered images of current and past market data, leading to more accurate predictions and reduced randomness.
Researchers demonstrate that locally observed robot distribution can correlate with environmental features, such as exits in office-like environments. This approach enables trapped office workers to navigate their way out of a collapsed building, even in scenarios where robots lack communication or sensors.
Researchers have created a distributed sensor fault diagnosis algorithm to detect and isolate multiple sensor faults in large-scale HVAC systems. The algorithm can be applied to both existing Building Management Systems and plug-in IoT systems, notifying users and operators about faulty measurements and sensor locations.
A group of researchers developed a new way for robots to pool data in real-time, allowing them to navigate difficult terrain as a team. The system uses a centralized data cloud, where each robot draws on data from other robots to steer clear of obstacles.
Researchers developed a more efficient automated parking guidance control strategy that mimics the approach to parallel parking commonly used by human drivers. This new method simplifies control rules and strategies, reducing computing and storage resources required in vehicles.
A team from Deakin University in Australia developed an improved sight-correcting system for self-driving vehicles. By watching human operators complete tasks, the vehicles can learn to make decisions based on visual information, reducing the need for extensive training data.
Researchers have developed a new optimization method for tracking self-driving car trajectories that reduces errors while keeping computation demands low. The method prioritizes passenger comfort, aiming to replicate human drivers' ability to balance speed, safety, and route choice.
A team of researchers has developed a novel attack detection scheme that captures vulnerable communication links, allowing systems to react optimally even under attack. The approach uses an adaptive observer to detect the onset of attacks and learns how they disrupt the system, enabling it to perform better under duress.
Researchers developed a control method that allows robots to better lift and move patients without compensating for friction, improving patient safety and comfort. The next step is to add a torso to the robot's arm, making it more human-like.
A new AI-powered method predicts building energy consumption by analyzing environmental and operational parameters, offering a significant improvement in energy management. The hybrid deep learning model has the potential to be applied in various smart buildings and cities.
Researchers propose a new computational analysis method that uses keystroke time series data to detect early stages of Parkinson's disease. The method, based on fuzzy recurrence plots, is less physically demanding than traditional testing methods and shows encouraging results.