Adaptive Scaling Factor Research Articles

Adaptive Perturbation for Adversarial Attack.

In recent years, the security of deep learning models achieves more and more attentions with the rapid development of neural networks, which are vulnerable to adversarial examples. Almost all existing gradient-based attack methods use the sign function in the generation to meet the requirement of perturbation budget on L∞ norm. However, we find that the sign function may be improper for generating adversarial examples since it modifies the exact gradient direction. Instead of using the sign function, we propose to directly utilize the exact gradient direction with a scaling factor for generating adversarial perturbations, which improves the attack success rates of adversarial examples even with fewer perturbations. At the same time, we also theoretically prove that this method can achieve better black-box transferability. Moreover, considering that the best scaling factor varies across different images, we propose an adaptive scaling factor generator to seek an appropriate scaling factor for each image, which avoids the computational cost for manually searching the scaling factor. Our method can be integrated with almost all existing gradient-based attack methods to further improve their attack success rates. Extensive experiments on the CIFAR10 and ImageNet datasets show that our method exhibits higher transferability and outperforms the state-of-the-art methods.

A modified fifth‐order WENO‐Z scheme based on the weights of the reformulated adaptive order WENO scheme

AbstractA modified fifth‐order WENO‐Z scheme is proposed by analogy with the non‐normalized weights of the reformulated fifth‐order adaptive order (AO) WENO scheme. We show that if the original fifth‐order WENO‐AO scheme is rewritten as the form of the conventional WENO combination, the resulting non‐normalized weights can be divided into three parts: a constant one term, a local stencil smoothness measure term and a global stencil smoothness measure term. In order to make use of the latter two terms for constructing a modified WENO‐Z scheme with enhanced performance, we change the form of the third term and introduce an adaptive scaling factor to adjust the contributions from the second and third terms. Numerical examples show that the modified fifth‐order WENO‐Z scheme has the advantage of high resolution in smooth regions and sharp capturing of discontinuities, and it can obtain evidently better results for shocked flows with small‐scale structures compared with the recently developed WENO‐Z+ and WENO‐Z+M schemes.

Differential evolutionary using opposition-based learning for multimodal multi-objective optimization

The correspondence between the decision space and the objective space is often many-to-one in multi-objective optimization problems. Therefore, a class of problems with such mapping relationships is defined as a MMOPs. For these problems, how to ensure the final solution converges to each Pareto solution set and guarantees the diversity of the algorithm is an urgent problem. The method of the paper with opposition-based strategy, a multimodal multi-objective optimization algorithm, is proposed. The algorithm proposed is called MMODE_OP, which is framed by a differential evolutionary algorithm, and opposition-based learning is applied to the initialization phase and generation-hopping phase to filter out the more promising individuals in the population for iteration to enhance the global search capability and the diversity of population. In addition, different Gaussian perturbation strategies are adopted with iteration to achieve the search of the neighborhood, which can further not only improve the quality of the Pareto solution set but also enable the convergence of the Pareto solution set quickly. This method improves the algorithm’s local and global search ability, and enables multiple the Pareto solution set and improving the convergence. In the meantime, adaptive scaling factors and crossover factors are designed in this paper to enhance the improved search capability. Finally, the experiment results of MMODE_OP and other excellent algorithms on 13 test problems corroborate the proposed methods have superior performance.

ESF-YOLO: an accurate and universal object detector based on neural networks.

As an excellent single-stage object detector based on neural networks, YOLOv5 has found extensive applications in the industrial domain; however, it still exhibits certain design limitations. To address these issues, this paper proposes Efficient Scale Fusion YOLO (ESF-YOLO). Firstly, the Multi-Sampling Conv Module (MSCM) is designed, which enhances the backbone network's learning capability for low-level features through multi-scale receptive fields and cross-scale feature fusion. Secondly, to tackle occlusion issues, a new Block-wise Channel Attention Module (BCAM) is designed, assigning greater weights to channels corresponding to critical information. Next, a lightweight Decoupled Head (LD-Head) is devised. Additionally, the loss function is redesigned to address asynchrony between labels and confidences, alleviating the imbalance between positive and negative samples during the neural network training. Finally, an adaptive scale factor for Intersection over Union (IoU) calculation is innovatively proposed, adjusting bounding box sizes adaptively to accommodate targets of different sizes in the dataset. Experimental results on the SODA10M and CBIA8K datasets demonstrate that ESF-YOLO increases Average Precision at 0.50 IoU (AP50) by 3.93 and 2.24%, Average Precision at 0.75 IoU (AP75) by 4.77 and 4.85%, and mean Average Precision (mAP) by 4 and 5.39%, respectively, validating the model's broad applicability.

Adaptive Orthogonal Basis Function Detection Method for Unknown Magnetic Target Motion State

Traditional methods of orthogonal basis function decomposition have been extensively used to detect magnetic anomaly signals. However, the determination of the relative velocity between the detection platform and the magnetic target remains elusive in practical detection. And, the non-ideal uniform motion of the magnetic target further complicates the process and adversely impacts the detector’s performance. To address this challenge, this paper introduces an adaptive scale factor method based on orthogonal basis function decomposition. This new method can be used to adjust the relative velocity between the detection platform and the magnetic target and to refine the characteristic time in the basis functions. In this paper, a mathematical relationship between the scale factor and the relative velocity is established, which is subsequently fitted into a Gauss function curve. The optimal scale factor, denoted as β, is adaptively chosen from the fitting curve when the magnetic target moves at a non-ideal uniform velocity with an unknown motion state. The results of simulations indicate that the scale factor improves the signal-to-noise ratio of the magnetic anomaly signals in a non-ideal state. And, this method can improve the energy value of OBF decomposition by 17.7%. Simultaneously, this method ensures that the moment the magnetic target passes the CPA coincides with the energy peak of the orthogonal basis detection, which improves the accuracy by 54.1% compared with the traditional method. The effectiveness and precision of the proposed method are verified using simulations and practical experiments.

A hypervolume-based cuckoo search algorithm with enhanced diversity and adaptive scaling factor

Existing multi-objective cuckoo search (MOCS) algorithms are based on either Pareto dominance or decomposition. However, when dealing with complex multi-objective problems (MOPs), the Pareto dominance-based algorithms face a decrease in selection pressure, and the decomposition-based algorithms easily gain poor distributions. The objective of this paper is to repurpose an indicator-based MOCS by combining improved diversity enhancement (IDE) and adaptive scaling factor (ASF) for MOPs. In the proposed algorithm, hypervolume is used as the indicator to guarantee better convergence and enough spread of the population. IDE chooses the large hypervolume to rebuild the parent population to compensate for the lack of population diversity. Additionally, ASF makes full use of individuals information to enhance the search ability of Lévy component in cuckoo search. Comprehensive experiments on 31 benchmark functions including two classical suites ZDT, WFG, and one challenged suite proposed in CEC2019, as well as 8 real-world problems were conducted to test the proposed algorithm. Compared with several state-of-the-art multi-objective evolutionary algorithms, the effectiveness and efficiency of our proposed method were demonstrated by the results.

Sky-view images aided NLOS detection and suppression for tightly coupled GNSS/INS system in urban canyon areas

Although the global navigation satellite system (GNSS) and inertial navigation system (INS) integration framework has many advantages compared to the standalone GNSS and INS systems, its accuracy can be severely degraded in urban canyon areas, due to non-line-of-sight (NLOS) signals. Detection and suppression of NLOS signals are the key to improving the location accuracy of urban areas. Therefore, this paper proposes a method for NLOS detection using the improved region growing algorithm on sky-view images captured by a fish-eye camera. After obtaining NLOS detection results, this paper constructs a weighted model with an adaptive scale factor to suppress the influence of NLOS signals. Experimental results show that the proposed method can effectively improve the performance of NLOS detection and suppression. On this basis, a tightly coupled GNSS/INS integration system based on an extended Kalman filter is developed and tested on the vehicle equipment. Results show that the proposed method outperforms the traditional GNSS single point positioning (SPP) and tightly coupled GNSS SPP/INS integration on positioning accuracy. The root mean square error can be reduced by 33.7% and 24.6% in the north and east directions respectively. It indicates that the tightly coupled GNSS SPP/INS integration aided by the fish-eye camera has certain research value and potential on positioning in urban canyon areas.

Secure and optimized satellite image sharing based on chaotic eπ map and Racah moments

The objective of this paper is to contribute to the sharing, transfer, and secured indexing of large color images. Watermarking and digital encryption have emerged as alternative and complementary solutions to ensure authorized access, facilitate authentication of content, or prevent illegal reproduction. For this reason, we suggest a new system based on four techniques derived from various transformation and decomposition methods, such as the discrete orthogonal Racah moment transformations (DORMT), which allows the information capacity to be reduced to a number of coefficients, as well as the discrete wavelet transform (DWT), which makes it possible to decompose into a certain number of multilevel sub-bands, to have the sub-band that represents most of the image information. In addition, Schur decomposition (SD) and singular value decomposition (SVD) have been used to increase security. The proposed system consists of two main phases. The first phase is the integration of a color watermark, first applying DORMT-SD-SVD to a color host image and then applying DWT-SD-SVD to achieve a compromise between capacity, invisibility and robustness in terms of quality and privacy. The second phase is the encryption of the watermarked image with an eπ map to increase security for sharing, transfer and indexing. This system uses an Aquila Optimization Algorithm (AOA) as an optimization procedure to find the best set of DORMT parameters and provide a dynamic and adaptive scale factor. The experimental results, in which the proposed system's properties are tested and compared to those of other systems, make it clear that the proposed system is useful in terms of its level of security, capacity, invisibility, and resistance to attacks that use signal processing.

Analysis and enhancement of MPPT technique to increase accuracy and speed in photovoltaic systems under different conditions

There has been tremendous growth in the use of renewable energy sources (RES) over the last few years. Solar systems, because of their unique properties, such as being noiseless, eco-friendly, and free of pollutants, are the most feasible source of energy among all RESs. However, a major drawback of photovoltaic (PV) systems is their low energy conversion efficiency, unstable and intermittent nature. In order to extract electrical energy from solar panels, it is necessary to connect them to the power system through DC-DC converters. As a result, maximum power point tracking (MPPT) algorithms are necessary in order to increase the amount of power harvested from solar PV systems. Accordingly, a method is presented in this paper for the design of an MPPT control approach for PV systems using fuzzy gain scheduling for proportional-integral-derivative (FGS-PID) controllers coupled with adaptive scaling factors for fuzzy input signals. The developed adaptive FGS-PID controller uses a two-level control system structure that incorporates the advantages of fuzzy logic and conventional PID control. Additionally, to fine-tune the member functions (MFs) of FGS, an integration of the fuzzy logic controller (FLC) and the modified fluid search optimization (MFSO) algorithm is employed. In order to evaluate the efficiency of the developed control systems, MATLAB software has been chosen for the simulations. The developed approach has proven highly effective in automatically regulating the triangular MFs of inputs and outputs and has shown superior performance. One of the key benefits of the proposed technique is that it enables fast and accurate tracking of the PV system's highest generated power in a short period of time.

Research on Green Vehicle Path Planning of AGVs with Simultaneous Pickup and Delivery in Intelligent Workshop

In this study, we present and discuss a variant of the classic vehicle routing problem (VRP), the green automated guided vehicle (AGV) routing problem, which involves simultaneous pickup and delivery with time windows (GVRPSPDTW) in an intelligent workshop. The research object is AGV energy consumption. First, we conduct a comprehensive analysis of the mechanical forces present during AGV transportation and evaluate the overall operational efficiency of the workshop. Then, we construct a green vehicle path planning model to minimize the energy consumption during AGV transportation and the standby period. Hence, the problems considered in this study are modeled based on asymmetry, making the problem solving more complex. We also design a hybrid differential evolution algorithm based on large neighborhood search (DE-LNS) to increase the local search ability of the algorithm. To enhance the optimal quality of solutions, we design an adaptive scaling factor and use the squirrel migration operator to optimize the population. Last, extensive computational experiments, which are generated from the VRPSPDTW instances set and a real case of an intelligent workshop, are designed to evaluate and demonstrate the efficiency and effectiveness of the proposed model and algorithm. The experimental results show that DE-LNS yields competitive results, compared to advanced heuristic algorithms. The effectiveness and applicability of the proposed algorithm are verified. Additionally, the proposed model demonstrates significant energy-saving potential in workshop logistics. It can optimize energy consumption by 15.3% compared with the traditional VRPSPDTW model. Consequently, the model proposed in this research carries substantial implications for minimizing total energy consumption costs and exhibits promising prospects for real-world application in intelligent workshops.

A Novel Color Image Watermarking Method with Adaptive Scaling Factor Using Similarity-Based Edge Region

This study aimed to deal with three challenges: robustness, imperceptibility, and capacity in the image watermarking field. To reach a high capacity, a novel similarity-based edge detection algorithm was developed that finds more edge points than traditional techniques. The colored watermark image was created by inserting a randomly generated message on the edge points detected by this algorithm. To ensure robustness and imperceptibility, watermark and cover images were combined in the high-frequency subbands using Discrete Wavelet Transform and Singular Value Decomposition. In the watermarking stage, the watermark image was weighted by the adaptive scaling factor calculated by the standard deviation of the similarity image. According to the results, the proposed edge-based color image watermarking technique has achieved high payload capacity, imperceptibility, and robustness to all attacks. In addition, the highest performance values were obtained against rotation attack, to which sufficient robustness has not been reached in the related studies.

Instance-Aware Deep Graph Learning for Multi-Label Classification

Graph convolutional neural network (GCN) has effectively boosted the multi-label image recognition task by modeling correlation among labels. In previous methods, label correlation is computed based on statistical information through label diffusion, and therefore the same for all samples. This, however, makes graph inference on labels insufficient to handle huge variations among numerous image instances. In this paper, we propose an instance-aware graph convolutional neural network (IA_GCN) framework for the multi-label classification. As a whole, two fused branches of sub-networks are involved in the framework: a global branch modeling the whole image and a local branch exploring dependencies among regions of interests (ROIs). For both the branches, an image-dependent label correlation matrix (ID_LCM), fusing both the statistical label correlation matrix (LCM) and an individual one of each image instance, is constructed to inject adaptive information of label-awareness into the learned features of the model through graph convolution. Specifically, the individual LCM of each image is obtained by mining the label dependencies based on the predicted label scores of those detected ROIs. In this process, considering the contribution differences of ROIs to multi-label classification, variational inference is introduced to learn adaptive scaling factors for those ROIs by considering their complex distribution. Finally, extensive experiments on MS-COCO and VOC datasets show that our proposed approach outperforms existing state-of-the-art methods.

Dynamic Characteristic Monitoring of Wind Turbine Structure Using Smartphone and Optical Flow Method

The dynamic characteristics of existing wind turbine structures are usually monitored using contact sensors, which is not only expensive but also time-consuming and laborious to install. Recently, computer vision technology has developed rapidly, and monitoring methods based on cameras and UAVs (unmanned aerial vehicles) have been widely used. However, the high cost of UAVs and cameras make it difficult to widely use them. To address this problem, a target-free dynamic characteristic monitoring method for wind turbine structures using portable smartphone and optical flow method is proposed by combining optical flow method with robust corner feature extraction in ROI (region of interest). Firstly, the ROI region clipping technology is introduced after the structural vibration video shooting, and the threshold value is set in the ROI to obtain robust corner features. The sub-pixel displacement monitoring is realized by combining the optical flow method. Secondly, through three common smartphone shooting state to monitor the structural displacement, the method of high pass filtering combined with adaptive scaling factor is used to effectively eliminate the displacement drift caused by the two shooting states of standing and slightly walking, which can meet the requirements of structural dynamic characteristics monitoring. After that, the structural displacement is monitored by assembling the telephoto lens on the smartphone. The accuracy of displacement monitored by assembling the telephoto lens on the smartphone is investigated. Finally, the proposed monitoring method is verified by the shaking table test of the wind turbine structure. The results show that the optical flow method, combined with smartphones, can accurately identify the dynamic characteristics of the wind turbine structure, and the smartphone equipped with a telephoto lens is more conducive to achieving low-cost wind turbine structure dynamic characteristics monitoring. This research can provide a reference for evaluating the condition of wind turbine structures.

An efficient adaptive scaling factor for 4×4 DCT image watermarking

An efficient adaptive scaling factor for 4×4 DCT image watermarking

Finite-Time Tracking Control Based on Immersion and Invariance with Dynamically Scaling Factor for Agile Missiles

A novel finite-time tracking control algorithm with disturbances observer is investigated for agile missiles in the presence of mismatched and matched disturbances. A finite-time disturbance observer with the continuous super-twisting algorithm is designed to quickly estimate the mismatched and matched disturbances. An adaptive law on the basis of immersion and invariance theory (I&I) is proposed to estimate the uncertainty of aerodynamics and compensate for the inaccuracy of modeling. An adaptive dynamic scaling factor supervised by a designed supervision factor is implemented. Moreover, a novel sliding mode controller is designed, and a barrier Lyapunov function (CTV-BLF) containing the system state constraints is constructed, which can guarantee that the violation of the constraint will not appear. The finite-time stability of the new proposition is proven via Lyapunov-based analysis. Comparative simulation results illustrate the effectiveness of the proposed scheme.

Fault Diagnosis Analysis of Angle Grinder Based on ACD-DE and SVM Hybrid Algorithm

Due to the complex structure of the angle grinder and the existence of multiple rotating parts, the coupling phenomenon of the data results in the complexity and chaos of the data. The market scale of angle grinder is huge. Manual diagnosis and traditional diagnosis are difficult to meet the requirements, so a fault diagnosis method of angle grinder that is based on adaptive parameters and chaos theory of dual-strategy differential evolution algorithm (ACD-DE) and SVM model hybrid algorithm is proposed by combining a chaos-mapping algorithm, dynamic and adaptive scale factor, and crossover factor. The effectiveness and robustness of the algorithm are proven by solving eight test functions. The acceleration signal is decomposed by wavelet packet decomposition and reconstruction, and a variety of sensor signals are processed and constructed as feature vectors. The training set and the test set of the fault diagnosis model are divided. SVM model is used as the fault diagnosis model and optimized by ACD-DE. Based on the fault data of the angle grinder, the hybrid algorithm is compared with other optimization algorithms and other machine learning models; the comparison results show that the performance of the improved differential evolution algorithm is improved, in which the precision rate is 98.81%, the recall rate is 98.74%, and the F1 score is 0.9877. Experiments show that the hybrid algorithm has strong diagnosis accuracy and robustness.

An Improved Cuckoo Search Algorithm Based on Inertial Weight and Scaling Factor

<p>This paper presents a novel and an improved cuckoo search algorithm (ISCS), different from other techniques, which sets nonlinear decreasing inertia weight and adaptive scaling factor. As the iteration goes on, these two parameters are controlled by two functions to change the iterations dynamically. At the beginning of the iteration, the values of the two parameters are favorable for global search, and at the end of the iteration, their values are more favorable for local optimization. In this work, 23 classical benchmark functions are selected to improve the accuracy and convergence speed by conducting the simulation experiments on CS, ISCS and other three algorithms. The results show that the improved cuckoo algorithm enhances the accuracy of understanding and speeds up the convergence of the curves. Finally, ISCS is applied to probabilistic neural network (PPN) neural network classification and recognition technology, and the results show that the optimization of ISCS can effectively improve the classification accuracy of the test sets, and has diverse applications.</p> <p> </p>

Single image based 3D human pose estimation via uncertainty learning

In monocular image scenes, 3D human pose estimation exhibits inherent ambiguity due to the loss of depth information and occlusions. Simply regressing body joints with high uncertainties will lead to model overfitting and poor generalization.In this paper, we propose an uncertainty-based framework to jointly learn 3D human poses and the uncertainty of each joint. Our proposed joint estimation framework aims to mitigate the adverse effects of training samples with high uncertainties and facilitate the training procedure. To be specific, we model each body joint as a Laplace distribution for uncertainty representation. Since visual joints often exhibit low uncertainties while occluded ones have high uncertainties, we develop an adaptive scaling factor, named the uncertainty-aware scaling factor, to ease the network optimization in accordance with the joint uncertainties.By doing so, our network is able to converge faster and significantly reduce the adverse effects caused by those ambiguous joints. Furthermore, we present an uncertainty-aware graph convolutional network by exploiting the learned joint uncertainties and the relationships among joints to refine the initial joint localization. Extensive experiments on single-person (Human3.6M) and multi-person (MuCo-3DHP & MuPoTS-3D) 3D human pose estimation datasets demonstrate the effectiveness of our method.

Modified Multifidelity Surrogate Model Based on Radial Basis Function with Adaptive Scale Factor

Multifidelity surrogates (MFSs) replace computationally intensive models by synergistically combining information from different fidelity data with a significant improvement in modeling efficiency. In this paper, a modified MFS (MMFS) model based on a radial basis function (RBF) is proposed, in which two fidelities of information can be analyzed by adaptively obtaining the scale factor. In the MMFS, an RBF was employed to establish the low-fidelity model. The correlation matrix of the high-fidelity samples and corresponding low-fidelity responses were integrated into an expansion matrix to determine the scaling function parameters. The shape parameters of the basis function were optimized by minimizing the leave-one-out cross-validation error of the high-fidelity sample points. The performance of the MMFS was compared with those of other MFS models (MFS-RBF and cooperative RBF) and single-fidelity RBF using four benchmark test functions, by which the impacts of different high-fidelity sample sizes on the prediction accuracy were also analyzed. The sensitivity of the MMFS model to the randomness of the design of experiments (DoE) was investigated by repeating sampling plans with 20 different DoEs. Stress analysis of the steel plate is presented to highlight the prediction ability of the proposed MMFS model. This research proposes a new multifidelity modeling method that can fully use two fidelity sample sets, rapidly calculate model parameters, and exhibit good prediction accuracy and robustness.

Research on Optimal Control Method of Tennis Racket String Diameter Based on Kalman Filter Algorithm

In order to explore the optimal control method of tennis racket string diameter, this paper applies the Kalman filter algorithm to the collection and processing of tennis sports parameters. When the Kalman filter uses the optimal gain (the Kalman filter enters the steady state), the corresponding cost function is established based on the noncorrelated nature of its residual sequence, which is used as the judgment condition for the sampling strategy switch to improve the stability of the system. In addition, this paper improves the real-time performance and calculation accuracy of data transformation through adaptive sampling strategy and adaptive scale factor, improves the stability and estimation accuracy of the system as a whole, and builds an intelligent monitoring system. Finally, this paper systematically studied the optimization control method of tennis racket string diameter and verified that the Kalman filter algorithm can play a certain role in the optimization control of tennis racket string diameter.