The dc-dc multiphase series-capacitor (SC) buck converter is a promising single-stage candidate for efficiently stepping-down the increasingly common 48 V rack-level distribution bus voltage directly to the core volta...
The dc-dc multiphase series-capacitor (SC) buck converter is a promising single-stage candidate for efficiently stepping-down the increasingly common 48 V rack-level distribution bus voltage directly to the core voltage of emerging microprocessors. Although there are multiple benefits to increasing the SC buck inductor-count, the corresponding dynamic performance becomes increasingly degraded due to the decreasing inductor rising-current slew-rate. This paper introduces a topological modification to the conventional SC buck that adds a new higher voltage level to each inductor switching-node, where the conventional topology would otherwise only provide two levels. Derived from existing platform-level intermediate bus converters, this third voltage level increases the total combined inductor rising-current slew-rate, significantly improving the light-to-heavy load-transient response, all while maintaining the static-efficiency benefits of the SC buck. Minimum deviation and time-optimal transient response improvements are shown in simulation, and functionality is proven with a discrete 11-phase, 48V-to-1V prototype.
Deep learning (DL) has been extensively adopted in many applications, including disease prediction. Most DL-based applications are executed on a cloud server because the DL models are too large and complicated to be e...
详细信息
The accurate detection and segmentation of cells in microscopy image sequences play a crucial role in biomedical research and clinical diagnostic applications. However, accurately segmenting cells in low signal-To-noi...
详细信息
ISBN:
(纸本)9798350359527
The accurate detection and segmentation of cells in microscopy image sequences play a crucial role in biomedical research and clinical diagnostic applications. However, accurately segmenting cells in low signal-To-noise ratio images remains challenging due to dense touching cells and deforming cells with indistinct boundaries. To address these challenges, this paper investigates the effectiveness of marker-guided networks, including UNet, with Squeeze-And-Excitation (SE) or MixTransformer (MiT) backbone architectures. We explore their performance both independently and in conjunction with motion cues, aiming to enhance cell segmentation and detection in both real and simulated data. The squeeze and excitation blocks enable the network to recalibrate features, highlighting valuable ones while downplaying less relevant ones. In contrast, the transformer encoder doesn't require positional encoding, eliminating the need for interpolating positional codes, which can result in reduced performance when the testing resolution differs from the training data. We propose novel deep architectures, namely Motion USENet (MUSENet) and Motion UMiTNet (MUMiTNet), and adopt our previous method Motion UNet (MUNet), for robust cell segmentation and detection. Motion and change cues are computed through our tensor-based motion estimation and multi-modal background subtraction (BGS) modules. The proposed network was trained, tested, and evaluated on the Cell Tracking Challenge (CTC) dataset. When comparing UMiTNet to USENet, there is a noteworthy 23% enhancement in cell detection accuracy when trained on real data and tested on simulated data. Additionally, there is a substantial 32% improvement when trained on simulated data and tested on real data. Introducing motion cues (MUMiTNet) resulted in a significant 25% accuracy improvement over UMiTNet when trained on real data and tested on simulated data, and a 9% improvement when trained on simulated data and tested on real data. In the ge
Server less cloud computing has won vast recognition in current years because of its capability to provide green and price-effective solutions for numerous computing needs. On this model, cloud carriers manage the sca...
详细信息
This paper addresses the problem of quadrotor finite-time tracking control with saturated actuation inputs and unknown disturbances. In this article, the unknown disturbances act in both the translational and rotation...
详细信息
ISBN:
(数字)9798331542726
ISBN:
(纸本)9798331542733
This paper addresses the problem of quadrotor finite-time tracking control with saturated actuation inputs and unknown disturbances. In this article, the unknown disturbances act in both the translational and rotational motions of the under-actuated quadrotor. A nonlinear finite-time disturbance observer (NFTDO) is developed to handle the unknown compound terms of the external disturbances and the nonlinear effect of the actuators saturation. A hierarchical finite-time command-filtered backstepping control is investigated for the trajectory tracking of the quadrotor with actuator saturation. The finite-time command filters, virtual control signal, and error compensation mechanism are designed for the translational and the attitude dynamics, respectively. Such techniques allows the removal of the computation complexity problem incurred by the conventional hierarchical backstepping technique. The control design inevitably guarantees that the position and attitude tracking errors reach a sufficiently small region around the origin in finite time despite the presence of unknown disturbances. The stability analysis is presented using the Lyapunov stability theory, and the effectiveness of the proposed control strategy is demonstrated through numerical simulations.
Normally, utility-connected inverters are equipped with LCL filters to reduce the output current harmonics. But the system that is equipped with an LCL filter has a stability problem and requires a more complex contro...
详细信息
Solar energy is environmentally friendly energy and is available in large quantities. The energy produced is commonly converted into electricity by using solar panel (PV) systems for utilization. PV systems are genera...
详细信息
WiFi usability and scalability have contributed to an improvement in human living standards. However, this has also led to increased vulnerabilities and attack vectors within WiFi networks. Authentic users can face ne...
详细信息
ISBN:
(数字)9798331533038
ISBN:
(纸本)9798331533045
WiFi usability and scalability have contributed to an improvement in human living standards. However, this has also led to increased vulnerabilities and attack vectors within WiFi networks. Authentic users can face network failures and denial of service caused by attacks on wireless networks by some third-party. Wireless Intrusion Detection System (IDS) is needed to detect wireless attacks. Main of objective of wireless IDS is to monitor network traffic and classify if it is an attack or normal traffic. Modern anomaly based wireless intrusion detection systems use machine learning (ML) to learn from previous attacks in the dataset to learn patterns of the attack. Although this method is effective, but there is also downside to this approach. The time required to train and test these ML models are exceedingly high along with high computational costs. Big data is emerging technologies that is getting advanced day by day, due to which, network technology is increasing rapidly. For that reason, we need to discuss the issue, that is high computational costs. In our proposed solution, we used our custom oversampling technique BBOT (Balanced Boost Oversampling Technique) to rectify class imbalance, generating artificial samples for the classes, enhancing model performance. The results of the experiments show that the combination of feature selection, BBOT, and Decision Tree outperformed all other classifiers with reduced computational overhead, providing a practical and efficient solution for real-time intrusion detection in WiFi networks. When compared to XGBoost, the provided solution achieves a similar level of accuracy, but with significantly reduced training time.
A well-defined and well-designed experimentation process is crucial for meaningful research outcomes. This paper introduces a framework aimed at streamlining the experimentation process in networking testbeds, consist...
详细信息
ISBN:
(数字)9798350377644
ISBN:
(纸本)9798350377651
A well-defined and well-designed experimentation process is crucial for meaningful research outcomes. This paper introduces a framework aimed at streamlining the experimentation process in networking testbeds, consisting of five key components: i) the experiment driver, containing available experiment tools, ii) the control tools to enhance capabilities and create complex scenarios, iii) the monitoring implementation, providing researchers with meaningful insights into gathered metrics, iv)the database component, which will be the main repository for storing experimental results and v) the analytics component, which could provide insights on patterns and hidden knowledge found in the extracted data, with the help of algorithms and artificial intelligence. The framework can be applied to various networking testbeds and is followed by descriptions of two Proof of Concepts (PoCs) demonstrating its implementation.
暂无评论