Conducting large-scale numerical computations to obtain flow field during the hypersonic vehicle engineering design phase can be excessively costly. Although deep learning algorithms enable rapid prediction with high-precision, they require a significant investment in training samples, contradicting motivation of reducing cost acquiring field. The combination feature extraction and regression also achieve high-precision fields, which is more suitable tackle three-dimensional small dataset....

A Gaussian process regression (GPR) model based on an improved automatic kernel construction (AKC) algorithm using beam search is proposed to establish a surrogate between lift body shape parameters and aerodynamic coefficients with various training sets sizes. The precision of our assessed through tenfold cross-validation. AKC-GPR algorithm, polynomial regression, support vector (SVR) are employed construct the model. interpolation extrapolation capabilities model, as generated by examined...

In order to solve the problem of face recognition in complex environments being vulnerable illumination change, object rotation, occlusion, and so on, which leads imprecision target position, a algorithm with multi-feature fusion is proposed. This study presents new robust face-matching method named SR-CNN, combining rotation-invariant texture feature (RITF) vector, scale-invariant transform (SIFT) convolution neural network (CNN). Furthermore, graphics processing unit (GPU) used parallelize...

The issue of hypersonic boundary layer transition prediction is a critical aerodynamic concern that must be addressed during the design process high-speed vehicles. In this context, we propose an advanced mesoscopic method couples gas kinetic scheme (GKS) with Langtry–Menter model, including its three modification methods, tailored for accurate predictions flows. new incorporates turbulent energy term into Maxwellian velocity distribution function, and it effects modifications on within...

This paper investigates the problem of adaptive tracking control for quadcopter in presence nonlinear configuration uncertainties. It utilizes a real-time brain-inspired learning (RBiLC) method to address challenges posed by time-varying uncertain instructions. To issue flight law reconfiguration caused unknown changes fuselage (e.g., propellers or motors), this introduces an online learning-evaluation-optimization reconstruction mechanism based on RBiLC. The proposed controller mitigates...

Two problems exist in the study of trajectory optimization problem powered hypersonic gliding vehicles (HGVs) due to insufficient consideration overall design constraints as well strong couplings among relevant disciplines: (1) engine and thrust models are not compatible with existing HGV; (2) configuration parameters HGV included variables during (i.e., propulsion discipline is decoupled process design), thus failing fully explore effect power improve performance HGV. Therefore, application...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Federated graph learning (FedGL) is an emerging federated (FL) framework that extends FL to learn data from diverse sources. for non-graph has shown be vulnerable backdoor attacks, which inject a shared trigger into the training such trained backdoored model can predict testing containing as attacker desires. However, FedGL against attacks largely unexplored, and no effective defense exists. In this paper, we aim address significant deficiency. First, propose effective, stealthy, persistent...

Motivation: Determining the potential transcriptomic signatures driving cortical thinning in neuromyelitis optica spectrum disorder (NMOSD) and multiple sclerosis (MS). Goal(s): To compare atrophy patterns underlying molecular mechanisms between NMOSD MS. Approach: A partial least squares (PLS) regression model was used to associate CTh alteration profile with expression of genes from Allen Human Brain Atlas (AHBA) database, then Metascape analysis performed identify functional biological...

Federated Learning (FL) is a novel client-server distributed learning framework that can protect data privacy. However, recent works show FL vulnerable to poisoning attacks. Many defenses with robust aggregators (AGRs) are proposed mitigate the issue, but they all broken by advanced Very recently, some renewed AGRs designed, typically clipping or/and filtering strate-gies, and promising defense performance against In this paper, we these also carefully designed Specifically, observe breaking...

More and more hospitals are now deploying mul-tiple cameras in trauma room for a multi-perspective remote observation. but video surveillance system can cause privacy breach by showing storing sensitive information of patients staff. We use OpenPose which is state-of-the-art human body skeletons estimation framework to extract 18 key skeleton points. For preservation, we apply the image obfuscation techniques heads, also replace truth background. proposed head detection method based on 5...

Video surveillance in trauma room can improve the security, patient safety, healthcare standards, and hygiene monitoring, but video system cause privacy breach by showing storing sensitive information of patients staff. Head blurring protect privacy, we still able to know most actions people with blurred head. But when postures videos are complex or there many people, using real-time detection program detect blur heads is challenging, a high missed rate. OpenPose state-of-the-art human body...

This paper presents a study that demonstrates the potential of using finite element (FE) lung model constructed 4D-MRI (3D + time) for tracking tumor motion during respiratory cycle. A series volumetric images one cancer patient was acquired over time under free breathing and sorted into phases. FE with volume which is at full exhale phase. Displacement field from this initial to subsequent 3D volumes in phases were derived deformable image registration technique. displacement provides...

Abstract The importance measurement of nodes in directed weighted networks is great significance for theoretical research and practical application networks. Aiming at the network model actual networks, based on special transformation graphs tree roots, this paper proposes a comprehensive quantitative assessment method index network’s global properties, local attributes, location bidirectionally propagation. This transforms to be analyzed graph root, constructs bidirectional depth-affected...