Strong Robustness Research Articles

Image-based fire detection using an attention mechanism and pruned dense network transfer learning

Fire accidents continually threaten lives and property, posing a great risk to public safety. Therefore, timely, accurate fire smoke and flame detection technology provides a technical foundation for ensuring personnel and property safety. This paper proposes a deep transfer learning model that integrates attention mechanisms and pruning techniques into the DenseNet network (P-DenseNet-A-TL) for detecting fire smoke and flame targets. To reduce the computational complexity of the fire detection model, we simplified the DenseNet network structure. To improve the recognition accuracy of the fire detection model, we incorporated attention mechanisms including a channel attention module (CAM) and spatial attention module (SAM) into the pruned dense network structure. To expedite the model training process, we also introduced a transfer learning model. Furthermore, to enhance the generalization ability and robustness of the model, we created a large-scale fire smoke and flame dataset. The experimental results show that our model (P-DenseNet-A-TL) achieved a test accuracy of 99.06%, F1 score of 99.09%, area under the curve (AUC) of 0.97, and a detection speed of 756 frames per second (FPS). The comparison experimental results and ablation experimental results indicate that our method achieves high detection accuracy and efficiency. Additionally, it possesses strong generalization capability and robustness.

Model-free compensation learning control of asymmetric hysteretic systems with initial state learning

In this article, a model-free compensation learning control scheme is presented for asymmetric hysteretic systems to achieve high-precision output tracking, where the effects of asymmetric input hysteresis nonlinearities are denoted by the asymmetric Prandtl–Ishlinskii model (APIM). In the presented control scheme, the feedforward iterative learning technique is generalized to the control design of asymmetric hysteretic systems. To improve the robustness and dynamic performance of the existing feedforward learning control for hysteretic systems, both the feedback action and the differential term are utilized to design a discrete-time PD-P open-closed-loop learning control law for simultaneously compensating the asymmetric input hysteresis nonlinearities and the linear dynamics effects without constructing compensators based on their models. The initial state error of such a system is also considered, a modified initial state learning strategy is proposed to ensure the initial state error tends to a prescribed level with the increasing of iterations. By fully analyzing the properties of the APIM, the convergence conditions with respect to the input error, the state error, and the tracking error along the iteration domain are given. The simulation and experimental results are provided to demonstrate strong robustness and excellent tracking accuracy with the proposed model-free compensation learning control scheme.

Recognition of α and β Radiation Waveforms Based on Neural Networks

Alpha and beta radiation do not possess penetrating properties to the skin, but they can adhere to the surface of the human skin. Inhalation through the mouth and nose often poses a greater risk to the human body compared to gamma radiation. Traditional methods for identifying alpha and beta particles have some drawbacks, such as high requirements for equipment's signal-to-noise ratio and significant impact of noise on identification results. We propose the utilization of lightweight neural network models for alpha and beta particle identification. These models exhibit strong generalization and robustness, enhancing the ability to resist noise interference during the identification process. Additionally, their lightweight nature facilitates deployment on devices, thereby contributing to the prevention of nuclear proliferation to some extent.

Research on adaptive coupling trajectory tracking anti-swing control strategy for three-dimensional double-pendulum overhead crane

In the transportation process of a three-dimensional double-pendulum overhead crane, the system is significantly influenced by the pronounced coupling introduced by the double-pendulum effect, posing a considerable challenge for the development of effective anti-swing control strategies. Moreover, uncertainties in certain system parameters and errors in trolley positioning contribute to the complexity of anti-swing control strategy design. To address these practical issues, a control strategy is proposed: First, an S-shaped transport trajectory with minimal positioning error is introduced, incorporating more system parameters into the coupling signal design to enhance system coupling. Based on this, an error-coupled trajectory signal is introduced. Second, the error-coupled trajectory signal is integrated into the energy function, and leveraging adaptive principles, an adaptive coupled trajectory tracking anti-swing control strategy is proposed to estimate uncertain system parameters online. Subsequently, the asymptotic stability of the equilibrium point of the closed-loop system is verified using the Lyapunov techniques and the Barbalat lemma. Finally, through simulations and experiments, it is demonstrated that the proposed control strategy not only ensures precise positioning of the trolley and bridge but also effectively suppresses oscillations of the hook and load, exhibiting excellent control performance. Even in scenarios where system parameters undergo changes or external disturbances are introduced, the proposed control strategy exhibits strong robustness and holds significant practical potential.

A fixed-time nonsingular terminal sliding mode control based on radial basis function neural network for vehicle active front steering system

In our work, a novel fixed-time nonsingular terminal sliding mode controller is developed to improve the control performance for the active front steering system of electric vehicles. First, the sliding mode surface and the sliding mode convergence law are constructed so that the sliding variable can be stabilized to zero at a fixed time. Moreover, the radial basis function neural network is introduced to estimate the systematic uncertainties, which can make the system have stronger anti-disturbance ability. The Lyapunov analysis also is given to verify the fixed-time stability of the whole system. The presented control scheme possesses two appealing advantages, including the strong robustness to perturbations and the fixed-time convergence to zero. Notably, the settling time is independent of the initial value while only related to the controller parameters. Through the joint simulation of CarSim and MATLAB, the given controller is compared with the conventional sliding mode controller and nonsingular terminal sliding mode controller to demonstrate the plausibility and validity of the presented control strategy.

From PID to ADRC and back: Expressing error‐based active disturbance rejection control schemes as standard industrial 1DOF and 2DOF controllers

AbstractIn this paper, we uncover a new connection between standard proportional‐integral (PI), proportional‐integral‐derivative (PID) controllers and active disturbance rejection control (ADRC), from which we establish formal conditions of equivalence between the two control schemes. Using the equivalence, we devise a step‐by‐step procedure of transitioning from PI/PID to error‐based ADRC. We also show how to go from one degree‐of‐freedom (1DOF) to two degree‐of‐freedom (2DOF) ADRC while retaining a standard 2DOF PI/PID structure. Both procedures facilitate expressing error‐based ADRC schemes as standard industrial 1DOF and 2DOF controllers. This allows the designed controller to have the desired characteristic of ADRC (i.e., strong robustness against internal and external uncertainties) while still being expressed in a form that is familiar to industrial practitioners, where PI/PID structures are still the workhorse of modern control systems. The paper's results ensure backward compatibility of future ADRC‐based solutions and foster the adoption of active disturbance rejection‐based methods in industrial practice as a viable alternative to standard controllers. To further support the findings, a set of tests is conducted in time and frequency domain.

Inversion of Soil Moisture Content in Cotton Fields Using GBR-RF Algorithm Combined with Sentinel-2 Satellite Spectral Data

Soil moisture content plays a vital role in agricultural production, significantly influencing crop growth, development, and yield. Thoroughly understanding the specific soil moisture content in cotton fields is crucial for enhancing agricultural efficiency and driving sustainable agricultural development. This study utilized the gradient-boosting regression–random forest (GBR-RF) algorithm and the GBR and RF algorithms separately, in conjunction with Sentinel-2 satellite images, to estimate cotton soil moisture content, focusing on the B1–B8 bands and in particular the sensitive B6, B7, and B8 bands. The soil data in the jujube orchard of the study area were collected using soil augers at a depth of 30 cm, with soil data collected from a depth of 20 to 30 cm. The findings revealed that the integrated learning algorithm GBR-RF demonstrated high accuracy, with R2, MAE, and MSE results of 0.8838, 1.0121, and 1.6168, respectively. In comparison, the results using just the GBR algorithm yielded R2, MAE, and MSE values of 0.8158, 1.1327, and 1.9645, respectively, while those obtained from the RF algorithm were 0.8415, 1.0680, and 1.8331, respectively. These results indicate that the algorithms exhibited strong generalization, robustness, and accuracy, with GBR-RF outperforming GBR and RF by 8.34% and 5.03%, respectively, in combination with using the B1–B8 bands for inversion. Furthermore, utilizing the full-band data resulted in R2 values that were up to 24.27% higher than those of the individual bands, affirming the efficacy of band combinations for improved accuracy. This study’s demonstration of the positive impact of integrated learning algorithms on estimating cotton soil moisture content underscores the advantages of multi-band data combinations over single-band data, highlighting their ability to enhance accuracy without significantly impacting errors. Importantly, this study’s findings, while not limited to a single experimental field, have broad applicability in cotton precision agriculture, offering valuable insights for research on yield enhancement and agricultural efficiency.

Distributed active vibration control for helicopter based on diffusion collaboration

The active vibration control technology has been successfully applied to several helicopter types. However, with the increasing of control scale, traditional centralized control algorithms are experiencing significant increase of computational complexity and physical implementation challenging. To address this issue, a diffusion collaboration-based distributed Filtered-x Least Mean Square algorithm applied to active vibration control is proposed, drawing inspiration from the concept of data fusion in wireless sensor network. This algorithm distributes the computation load to each node, and constructs the active vibration control network topology of large-scale system by discarding the weak coupling secondary paths between nodes, achieving distributed active vibration control. In order to thoroughly validate the effectiveness and superiority of this algorithm, a helicopter fuselage model is designed as the research object. Firstly, the excellent vibration reduction performance of the proposed algorithm is confirmed through simulations. Subsequently, specialized node control units are developed, which utilize STM32 microcontroller as the processing unit. Further, a distributed control system is constructed based on multi-processor collaboration. Building on this foundation, a large-scale active vibration control experimental platform is established. Based on the platform, experiments are carried out, involving the 4-input 4-output system and the 8-input 8-output system. The experimental results demonstrate that under steady-state harmonic excitation, the proposed algorithm not only ensures control effectiveness but also reduces computational complexity by 50%, exhibiting faster convergence speed compared with traditional centralized algorithms. Under time-varying external excitation, the proposed algorithm demonstrates rapid tracking of vibration changes, with vibration amplitudes at all controlled points declining by over 94%, proving the strong robustness and adaptive capability of the algorithm.

Probing PT-Symmetry Breaking of Non-Hermitian Topological Photonic States via Strong Photon-Magnon Coupling.

Light-matter interaction is crucial to both understanding fundamental phenomena and developing versatile applications. Strong coupling, robustness, and controllability are the three most important aspects in realizing light-matter interactions. Topological and non-Hermitian photonics have provided frameworks for robustness and control flexibility, respectively. How to engineer the properties of the edge state such as photonic density of state by using non-Hermiticity while ensuring topological protection has not been fully studied. Here we construct a parity-time-symmetric dimerized photonic lattice and probe the spontaneous PT-symmetry breaking of the edge states by utilizing the strong coupling between the photonic mode and a spin ensemble. Our Letter presents an accurate and almost noninvasive approach for investigating non-Hermitian topological states, while also offering methodologies for the implementation and manipulation of topological light-matter interactions.

Fault-tolerant control of automatic carrier landing using direct lift control based on nonsingular terminal sliding mode and active disturbance rejection control

This paper presents a fault-tolerant control scheme, combining nonsingular terminal sliding mode and active disturbance rejection control (NTSM-ADRC) to address actuator failure and external disturbances during the automatic carrier landing process of a carrier-based aircraft using direct lift control (DLC). First, the carrier-based aircraft model, the carrier air-wake model, and the actuator fault model were established. Second, the NTSM-ADRC controller is designed, The unmodeled dynamics of the system, the air-wake disturbance, and the fault term are treated as total disturbances and estimated accurately by extended state observer (ESO). To improve the response characteristics of the controller, the nonlinear error feedback control law is designed by combining the NTSMC. The Lyapunov function is constructed to prove the stability of the closed-loop system. The controller is applied to the aircraft DLC channel, attitude auxiliary channel, and approach power compensation system. The DLC improves the performance of fixed-wing aircraft by directly generating high lift through the flaps to change the aircraft trajectory. Finally, the method is tested by introducing various types of actuator failures. Simulation results demonstrate that the designed longitudinal fault-tolerant carrier landing system exhibits strong robustness and fault tolerance, thereby improving the accuracy of aircraft landing trajectory tracking.

Robust photo-induced pure spin current in graphene–graphane superlattices

The photogalvanic effect has been demonstrated to be an effective method to generate pure spin current. However, obtaining robust pure spin current remains a big challenge as most of the photo-induced pure spin current is dependent on the light polarization/phase angle, photon energies, or the spin alignment of electrodes. In this paper, we present a scheme for obtaining robust pure spin current in zigzag graphene–graphane superlattices (ZGSLs). Through systemic first-principles calculations, we demonstrate that robust generation of pure spin current can occur in ZGSLs with varying widths due to the centrosymmetry of the system's geometric structure and the antiferromagnetic magnetic features. Moreover, the generation of pure spin current does not depend on the photon energy or the polarization/phase angles for both linearly and elliptically polarized light, exhibiting very strong robustness. Our study suggests that ZGSLs could be highly promising candidates for practically realizing pure spin current in spintronics experiments, which presents an avenue for using graphene and its derivatives in advanced electronic devices. Furthermore, considering the experimental advancements in graphene and graphene-like materials, our investigation presents a universally applicable methodology for the generation of robust pure spin currents within low-dimensional graphene-like systems.

Research on the Control Method of a Brushless DC Motor Based on Second-Order Active Disturbance Rejection Control

This research addresses the issues of weak anti-disturbance ability, fast response, and incompatibility of overshoot in the control process of brushless DC motors (BLDCs). A six-step commutation control method based on second-order active disturbance rejection control (ADRC) is derived following the analysis of the BLDC model and the mathematical model of ADRC. A control model of the BLDC using both PI and ADRC is constructed. Detailed comparative and quantitative analyses of the simulation results using PI and ADRC are conducted, focusing on the anti-load disturbance capabilities using the integrated square error (ISE), integrated time square error (ITSE), integrated absolute error (IAE), and integrated time absolute error (ITAE). Experimental testing on the STM32F4 controller is also carried out, analyzing four error integral criteria in depth. The results indicate that both the ADRC and PI control modes can track the target signal without overshooting, demonstrating strong anti-load disturbance ability and robustness at varying working speeds. In the BLDC control system, using the ADRC control method can achieve fast and non-overshoot tracking of target signals compared to the PI control method, and ADRC has stronger resistance to load disturbances.

Enhancing Insect Sound Classification Using Dual-Tower Network: A Fusion of Temporal and Spectral Feature Perception

In the modern field of biological pest control, especially in the realm of insect population monitoring, deep learning methods have made further advancements. However, due to the small size and elusive nature of insects, visual detection is often impractical. In this context, the recognition of insect sound features becomes crucial. In our study, we introduce a classification module called the “dual-frequency and spectral fusion module (DFSM)”, which enhances the performance of transfer learning models in audio classification tasks. Our approach combines the efficiency of EfficientNet with the hierarchical design of the Dual Towers, drawing inspiration from the way the insect neural system processes sound signals. This enables our model to effectively capture spectral features in insect sounds and form multiscale perceptions through inter-tower skip connections. Through detailed qualitative and quantitative evaluations, as well as comparisons with leading traditional insect sound recognition methods, we demonstrate the advantages of our approach in the field of insect sound classification. Our method achieves an accuracy of 80.26% on InsectSet32, surpassing existing state-of-the-art models by 3 percentage points. Additionally, we conducted generalization experiments using three classic audio datasets. The results indicate that DFSM exhibits strong robustness and wide applicability, with minimal performance variations even when handling different input features.

Finite-Time Adaptive Synchronization and Fixed-Time Synchronization of Fractional-Order Memristive Cellular Neural Networks with Time-Varying Delays

Asymptotic synchronization requires continuous external control of the system, which is unrealistic considering the cost of control. Adaptive control methods have strong robustness to uncertainties such as disturbances and unknowns. On the other hand, for finite-time synchronization, if the initial value of the system is unknown, the synchronization time of the finite-time synchronization cannot be estimated. This paper explores the finite-time adaptive synchronization (FTAS) and fixed-time synchronization (FDTS) of fractional-order memristive cellular neural networks (FMCNNs) with time-varying delays (TVD). Utilizing the properties and principles of fractional order, we introduce a novel lemma. Based on this lemma and various analysis techniques, we establish new criteria to guarantee FTAS and FDTS of FMCNNs with TVD through the implementation of a delay-dependent feedback controller and fractional-order adaptive controller. Additionally, we estimate the upper bound of the synchronization setting time. Finally, numerical simulations are conducted to confirm the validity of the finite-time and fixed-time stability theorems.

A Color- and Geometric-Feature-Based Approach for Denoising Three-Dimensional Cultural Relic Point Clouds.

In the acquisition process of 3D cultural relics, it is common to encounter noise. To facilitate the generation of high-quality 3D models, we propose an approach based on graph signal processing that combines color and geometric features to denoise the point cloud. We divide the 3D point cloud into patches based on self-similarity theory and create an appropriate underlying graph with a Markov property. The features of the vertices in the graph are represented using 3D coordinates, normal vectors, and color. We formulate the point cloud denoising problem as a maximum a posteriori (MAP) estimation problem and use a graph Laplacian regularization (GLR) prior to identifying the most probable noise-free point cloud. In the denoising process, we moderately simplify the 3D point to reduce the running time of the denoising algorithm. The experimental results demonstrate that our proposed approach outperforms five competing methods in both subjective and objective assessments. It requires fewer iterations and exhibits strong robustness, effectively removing noise from the surface of cultural relic point clouds while preserving fine-scale 3D features such as texture and ornamentation. This results in more realistic 3D representations of cultural relics.

Fault‐tolerant control of current residual vector three‐phase four‐switch motor drive system based on MLD model

AbstractAiming at the problem of single‐phase switch tube open circuit fault in inverters, a current residual vector three‐phase four‐switch fault‐tolerant control strategy based on the mixed logic dynamic (MLD) model of permanent magnet synchronous motor (PMSM) drive system is proposed. Firstly, the MLD and three‐phase four‐switch mathematical models of the motor drive system are established and a calculation method of DC bus capacitance is proposed. Secondly, to improve the dynamic performance of the motor, an improved vector control based on the MLD model is proposed. Finally, by setting the threshold and dividing the area of the current residual vector, the accuracy of inverter fault diagnosis is improved, and the fast and smooth switching of the drive system after the fault is realized. Simulation and experimental results show that this strategy can effectively reduce the influence of calculation error and system noise, and has strong robustness for the change of motor speed and load.

Sliding mode controller with neural network compensation for tank

The control effectiveness of the all-electric tank stabilizers directly affects the firing accuracy of the tank gun. In order to improve the firing accuracy of the moving tank, this paper proposes a sliding mode control (SMC) strategy using neural network feed-forward compensation for the tank bidirectional stabilizers. SMC technology can quickly and effectively deal with uncertainties, unmodeled terms, and external disturbances in complex tank systems. Neural networks possess the advantage of approximating arbitrary continuous functions in finite time, realizing the estimation of SMC control errors with feed-forward compensation. The Lyapunov stability analysis proves that the designed controller can achieve asymptotic stabilization in finite time. Finally, both co-simulation and physical experiments are conducted. The results show that the proposed control strategy possesses better tracking speed and tracking accuracy compared to the conventional controller and exhibits strong robustness.

Intelligent robust control for aviation electric fuel pump based on second order integral sliding mode with UDE

This paper proposes an original intelligent robust control method for a kind of aviation electric fuel pump (AEFP) systems, in order to realize on-demand fuel supply for aircraft engines, accurately and robustly. By considering the feature of AEFP, the intelligent robust control method is in a cascaded structure, which combines a novel second order integral sliding mode control (SOISMC) approach with uncertainty and disturbance estimator (UDE) strategy, together with a new multi-objective variable speed grey wolf optimization algorithm (MOVSGWO). To let the current loop of AEFP possess both rapidity and robustness, the SOISMC is presented based on an innovative second order integral sliding mode manifold. To realize that AEFP has strong robustness to both the changing of speed command and the pulsation in fuel pump, UDE is utilized for the speed loop of AEFP. The primary parameters in the intelligent robust controller, namely, UDE-SOISMC controller, are optimized by MOVSGWO. The MOVSGWO is a modified grey wolf optimization algorithm with variable speed by particle swarm optimization. Mismatched uncertainties and disturbances in AEFP can be handled by the proposed robust method. The calculation of instantaneous flow and dynamic torque in AEFP is illustrated. Simulation and experimental results verify the effectiveness of the UDE-SOISMC method.

The influence of gear load distribution based on coupled systems on gearbox meshing noise

Introduction: With the rapid development of the gearbox manufacturing industry, the internal gear response has received attention, and the control of meshing noise during gear operation has been studied. Conventional noise reduction methods are usually based on gear order, and with the improvement of gearbox manufacturing technology, this method gradually becomes difficult to cope with a wide range of data.Methods: To expand the search domain of noise control systems, this study combines gear response and gear order, and adds the condition of gear uniform load. For common noise reduction problems in composite systems, this study improves the time-varying stiffness excitation mechanism and generates a coupled system.Results: Finally, this study conducts experiments on the Gmnoi dataset and compares it with three systems including quantum genetics to verify the superiority of the proposed system. The suppression effects of the four systems on gear meshing noise were 98.4%, 95.8%, 93.5%, and 92.7%, respectively. Their highest performance for different gear groups was 623, 514, 406, and 423, respectively.Discussion: The experimental results showed that the proposed coupling system has strong robustness and high accuracy in controlling gearbox meshing noise, and is of great significance in reducing noise pollution and improving the working environment of the gearbox.

Bearing fault diagnosis under variable speed conditions on adaptive time frequency extraction mode decomposition

Improvements in measurement technology have made it possible to detect problems with rolling bearings more accurately, which is important to ensure that they work properly in mechanical systems under different variable speed conditions. Time–frequency distribution (TFD) methods are widely used in variable-speed rolling bearing fault diagnosis, we construct a new method: adaptive time frequency extraction mode decomposition (ATFEMD) by capturing the distinctive time–frequency information within the TFD through ridge extraction, subsequently, the reconstruction components are further refined into adaptive modes through the harmonic detection and noise testing process. This method is a time–frequency post-processing method that effectively solves the problems of time–frequency energy lack of concentration, poor robustness of instantaneous frequency extraction, and mode aliasing in signal decomposition. This article analyzes the simulated bearing vibration and test bench bearing vibration signals to demonstrate the performance of ATFEMD. Results indicated that the proposed method is characterized by strong robustness, and good feature extraction results compared to other methods.