Showing papers by "Zhiwei Gao published in 2022"

Class-Imbalance Privacy-Preserving Federated Learning for Decentralized Fault Diagnosis With Biometric Authentication

Abstract: Privacy protection as a major concern of the industrial big data enabling entities makes the massive safety-critical operation data of a wind turbine unable to exert its great value because of the threat of privacy leakage. How to improve the diagnostic accuracy of decentralized machines without data transfer remains an open issue; especially these machines are almost accompanied by skewed class distribution in the real industries. In this study, a class-imbalanced privacy-preserving federated learning framework for the fault diagnosis of a decentralized wind turbine is proposed. Specifically, a biometric authentication technique is first employed to ensure that only legitimate entities can access private data and defend against malicious attacks. Then, the federated learning with two privacy-enhancing techniques enables high potential privacy and security in low-trust systems. Then, a solely gradient-based self-monitor scheme is integrated to acknowledge the global imbalance information for class-imbalanced fault diagnosis. We leverage a real-world industrial wind turbine dataset to verify the effectiveness of the proposed framework. By comparison with five state-of-the-art approaches and two nonparametric tests, the superiority of the proposed framework in imbalanced classification is ascertained. An ablation study indicates that the proposed framework can maintain high diagnostic performance while enhancing privacy protection.

Relative Threshold-Based Event-Triggered Control for Nonlinear Constrained Systems With Application to Aircraft Wing Rock Motion

Abstract: This article concentrates on the event-driven controller design problem for a class of nonlinear single input single output parametric systems with full state constraints. A varying threshold for the triggering mechanism is exploited, which makes the communication more flexible. Moreover, from the viewpoint of energy conservation and consumption reduction, the system capability becomes better owing to the contribution of the proposed event-triggered mechanism. In the meantime, the developed control strategy can avoid the Zeno behavior since the lower bound of the sample time is provided. The considered plant is in a lower triangular form, in which the match condition is not satisfied. To ensure that all the states retain in a predefined region, a barrier Lyapunov function (BLF) based adaptive control law is developed. Due to the existence of the parametric uncertainties, an adaptive algorithm is presented as an estimated tool. All the signals appearing in the closed-loop systems are then proven to be bounded. Meanwhile, the output of the system can track a given signal as far as possible. In the end, the effectiveness of the proposed approach is validated by an aircraft wing rock motion system.

Failure-informed adaptive sampling for PINNs

Abstract: . Physics-informed neural networks (PINNs) have emerged as an effective technique for solving PDEs in a wide range of domains. Recent research has demonstrated, however, that the performance of PINNs can vary dramatically with different sampling procedures, and that using a fixed set of training points can be detrimental to the convergence of PINNs to the correct solution. In this paper, we present an adaptive approach termed failure-informed PINNs(FI-PINNs), which is inspired by the viewpoint of reliability analysis. The basic idea is to define a failure probability by using the residual, which represents the reliability of the PINNs. With the aim of placing more samples in the failure region and fewer samples in the safe region, FI-PINNs employs a failure-informed enrich-ment technique to incrementally add new collocation points to the training set adaptively. Using the new collocation points, the accuracy of the PINNs model is then improved. The failure probability, similar to classical adaptive finite element methods, acts as an error indicator that guides the refinement of the training set. When compared to the conventional PINNs method and the residual-based adaptive refinement method, the developed algorithm can significantly improve accuracy, especially for low regularity and high-dimensional problems. We prove rigorous bounds on the error incurred by the proposed FI-PINNs and illustrate its performance through several problems.

Robust Synchronized Data Acquisition for Biometric Authentication

Abstract: Owing to its unique, concealment and easy customization by combining different wrist and hand gestures, high-density surface electromyogram (HD-sEMG) is recognized as a potential solution to the next generation biometric authentication, which usually adopts a wireless body sensor network (BSN) to acquire the multi-channel HD-sEMG biosignals from distributed electrode arrays. For more accurate and reliable classification, biometric authentication requires the distributed biosignals to be sampled simultaneously and be well aligned, which means that the sampling jitters among the arrays need to be tiny. To synchronize data sampling clocks of a cluster of BSN nodes for biometric authentication, this article modifies the packet-coupled oscillators protocol by using a dynamic controller (D-PkCOs). This protocol only involves one-way single packet exchange, which reduces the communication overhead significantly. For the purpose of maintaining precise sampling of these BSN nodes subject to drifting clock frequency and varying delays, the dynamic controller is designed via the $H_\infty$ robust method, and it is proved that all the BSN nodes’ sampling jitters are bounded. The experimental results demonstrate that the D-PkCOs protocol can keep the sampling jitters less than a microsecond in a 10-node IEEE 802.15.4 network. The application of D-PkCOs to the BSN shows that the HD-sEMG signal with a high signal-to-noise ratio is obtained, which leads to better gesture classification performance.

Non-flocking and flocking for the Cucker-Smale model with distributed time delays

Abstract: In this paper, we study a flocking behavior that may or not appear for Cucker - Smale model with distributed time delays. For the short range communicated Cucker - Smale model, the flocking condition has strong restrictions on initial data. For this case, we mainly consider the non - flocking behavior. By establishing and appropriately estimating an inequality of the position variance such that the second order space moment is unbounded, we drive a sufficient condition for the non - existence of the asymptotic flocking when the time delays satisfy a suitable smallness assumption. Furthermore, we also provide a sufficient condition of asymptotic flocking. Finally, we present numerical simulations to validate the theoretical results.

Finite-Time Annular Domain Stability and Stabilization of Stochastic Systems With Semi-Markovian Switching

Abstract: In this paper, finite-time annular domain stability and stabilization of Itô stochastic systems with semi-Markovian switching are investigated, where the switching frequency is timevarying, and satisfies a unfixed polytope. A new inequality called reverse differential Gronwall inequality is proposed to address less conservative sufficient conditions for testing the finitetime annular domain stability, and its superiority to modified Gronwall inequality is analyzed. Moreover, sufficient conditions for the existence of state feedback and observer-based finitetime annular domain stabilizing controllers are obtained. In the sequel, a $\mathscr {L}$ × $N$ -mode algorithm is presented for bridging the relationship between the adjustable parameters and the range of transition rates. Finally, an example is provided to illustrate the effectiveness of the proposed methods.

Cloud-Based Event-Triggered Predictive Control for Heterogeneous NMASs Under Both DoS Attacks and Transmission Delays

Abstract: A novel compensation control method for heterogeneous multiagent systems under Denial-of-Service (DoS) attacks and transmission delays is investigated in this article. This control method has all the advantages of the cloud-based computation strategy, the adaptive event-triggered strategy, and the predictive control scheme. The adaptive event-triggering mechanism can adjust the event numbers adaptively, the predictive control can reduce or eliminate the negative effects brought out by both DoS attacks and transmission delays actively, while the cloud-based computation strategy can eliminate the negative effects completely as the same as there are no DoS attacks and transmission delays. Through the interval decomposition skill and the augmented system modeling method, the compensated geschlossenes system model is established. Moreover, the joint design for the feedback gain matrices and the event-triggered parameters is implemented. In the simulation part, five VTOL aircraft are used to demonstrate the theoretical results.

Fault Diagnosis and Fault-Tolerant Control of Energy Maximization for Wave Energy Converters

Abstract: This paper investigates a new wave energy converter (WEC) control problem, which is the energy maximization subject to sensor and actuator faults. Unexpected deviation of system variables from standard conditions, defined as a fault, degrades the control performance and even introduces damages breaking down the whole system. Fault detection for wave energy converters is therefore of great importance in maintaining the high reliability of the system. This paper presents a robust fault diagnosis approach effectively detecting sensor and actuator faults in real-time. A compensator is then designed to minimize the influence from faults and maintain the control performance. A non-causal linear optimal control is applied to maximize the energy output, in which the future excitation force is incorporated to determine the current control action. This approach can also be straightforwardly applied to other control methods. The parameters of the proposed fault detection method and fault-tolerant control method can be calculated off-line, which enhances the real-time implementation with a low computational burden. A realistic sea wave collected from the coast of Cornwall, U.K. is used to demonstrate the efficacy of the proposed approach.

Mean‐square strong stability and stabilization of discrete‐time stochastic systems with multiplicative noises

Abstract: This article investigates the mean‐square strong stability and stabilization of a discrete‐time stochastic system corrupted by multiplicative noises. First, the definition of the mean‐square (MS) strong stability is addressed to avoid overshoots in system dynamics, and two necessary and sufficient conditions for the MS‐strong stability are derived. Moreover, the relationship between MS‐strong stability and MS‐stability is given. Second, some necessary and sufficient conditions of the MS‐strong stabilization via state feedback (SF) and output feedback are obtained, respectively. Furthermore, analytical expressions of SF controller and static output feedback (SOF) controller are proposed, respectively. Finally, an equivalent design method for SOF controller and dynamic output feedback controller is presented.

A Deep Learning-Based Fault Diagnosis of Leader-Following Systems

Abstract: This paper develops a multisensor data fusion-based deep learning algorithm to locate and classify faults in a leader-following multiagent system. First, sequences of one-dimensional data collected from multiple sensors of followers are fused into a two-dimensional image. Then, the image is employed to train a convolution neural network with a batch normalisation layer. The trained network can locate and classify three typical fault types: the actuator limitation fault, the sensor failure and the communication failure. Moreover, faults can exist in both leaders and followers, and the faults in leaders can be identified through data from followers, indicating that the developed deep learning fault diagnosis is distributed. The effectiveness of the deep learning-based fault diagnosis algorithm is demonstrated via Quanser Servo 2 rotating inverted pendulums with a leader-follower protocol. From the experimental results, the fault classification accuracy can reach 98.9%.

Guest Editorial: Biometrics in Industry 4.0: Open Challenges and Future Perspectives

Abstract: THE paradigm of Industry 4.0 aims at reinventing procedures and business strategies in all industrial/factory fields. On the other hand, such a renovation generates massive amounts of data, lying from traditional tracking information of goods and services but also source codes, technology information, and intellectual property. This represents a serious threat for corporations that must invest, more than in the past, in protocols and procedures to keep their data secure from being stolen or widespread over networks. Therefore, this special section is devoted to selected papers focused on open challenges in the context of security and safety of Biometrics in Industry 4.0. Strong authentication mechanisms represent the key to ensure such a high level of security of data and biometrics may play a crucial role in this scenario. Biometric traits, both physical and behavioral, can be used to build up strong authentication platforms and infrastructure of trust enabling big corporations to protect their data against criminal attacks. On the other hand, the use of authentication mechanisms based on biometric solutions can often collide with national and international regulations which tend to protect sensitive information. While the use of biometrics for security can be developed in full awareness of the users, in case of safety the collection of biometric data concerns a more subtle concept of privacy. In a private context, the security aims at guaranteeing controlled access/use to spaces and devices for a limited number of users. In public, is the safety to be guaranteed, involving a wider number of users whose biometrics data need to be acquired and protected. The ethical and practical implication of the use of biometrics data in safety needs to be further explored. The choice of the biometric traits (either hard or soft) as well as the mechanism to put in place to protect those data and the user’s privacy ask for meticulous studies of the minimum number of traits to be involved in the monitored environment to achieve systems that are privacy preserving as well as robust and reliable. Augmented reality applications have been intensively promoted by Industry 4.0 but their usage in operative environments is not free from risks for employees and trainees. The analysis of behavioral biometric traits, and soft biometrics in general, can provide useful estimations on postures and gestures during augmented reality training sessions. They can be used to assess the suitability of the augmented reality application and indicate excessive physical workload that may attempt personnel safety.

Non-rechargeable battery remaining useful life prediction with interactive attention sequence to sequence network

Abstract: Non-rechargeable batteries remain as the main source of energy for small systems, owing to their unique advantages in energy density, safety, reliability and sustainability. Accurate prediction of the remaining useful life of the battery is not only beneficial to maintenance and production safety, but also can be regarded as a starting point for possible secondary life applications. In this study, an interactive attention sequence-to-sequence network is proposed for the remaining useful life prediction of the non-rechargeable batteries. The proposed approach can effectively extract the degenerate information of each variable-length sequence and dynamically weight the sequence features of different dimensions. For illustration, a case of primary battery dataset collected from the power supply system of 139 vibration sensors is utilized. The extensive experiments verify the effectiveness of the proposed approach.

Data-Driven Gap Metric Fault Diagnosis Technology With Its Application to DC-DC Converter

Abstract: This paper investigates a data-driven gap metric fault detection and isolation method for buck DC-DC converters with component faults. First, the averaged state space model of a buck DC-DC converter and its component (inductor, capacitor and load resistance) fault models are established. Second, a data-driven gap metric using subspace identification is utilized to detect the occurred component faults. Third, to isolate these faults, the concept of fault cluster is firstly developed and then the definition of fault isolation under gap matric is proposed. Based on it, a fault isolation condition is presented by solving its fault cluster center model and radius. Finally, the simulation and experiment are reported to demonstrate the effectiveness of the used method.

Quantized Adaptive Fault-like Model-based SMC Design for Vehicular Platoon with Actuator Deadzone and Distance Restrictions

Abstract: This paper aims to design an adaptive fault-like model-based sliding-mode control (SMC) law for a class of vehicle platoon systems, where the follower vehicles have quantized input, unknown actuator deadzone, and distance restrictions. Using adaptive terminal sliding mode control method and barrier Lyapunov function (BLF), an adaptive controller is designed to ensure the collision avoidance and communication connection keeping. In addition, the problem of the unknown actuator deadzone and input quantization is first transformed to a fault-like problem, which will simplify the control design procedure. Finally, the effectiveness of the proposed control strategy is verified by a numerical example.

ICoCTA 2021 Preface

Abstract: Control theory and engineering have witnessed dramatic achievements, which have made possible space travel and communication satellites, have assisted in the design of safe and efficient aircraft, ships, trains, and cars, have helped in developing cleaner chemical processes while addressing environmental concerns. The International Conference on Control Theory and Applications 2021(ICoCTA 2021), organized by the Hong Kong Society of Robotics and Automation (HKSRA), was scheduled to be held in Chengdu, China, but was held virtually on November 4-6, 2021 due to the Covid-19 pandemic. ICoCTA 2021 aims to bring together researchers, engineers, scientists and industry professionals in a unique platform and present their stimulating research and knowledge transfer ideas in Control Theory and Applications. The conference included two full days of technical sessions consisting of keynote lectures, oral presentations, and poster presentations. We were delighted and honored to invited Prof. Romeo Ortega (IEEE Fellow, Full Professor at ITAM, Mexico), Prof. Petros Ioannou (A.V. ‘BAL’ Chair Professor, University of Southern California) and Prof. YangQuan Chen (University of California, Merced) to share their insightful ideas and mind-blowing research as keynote speakers. After a couple of months well preparation and hard work, the proceeding of the International Conference on Control Theory and Applications 2021 (ICoCTA 2021) is smoothly published. On behalf of the conference organizing committee, we’d like to show our congratulation and sincere gratitude to all authors, peer reviewers, speakers, oral and poster presenters and listeners. We deeply appreciate their endeavor in making the conference an impressive success. List of ICoCTA Committee Members are available in the pdf

Sensor Fault Estimator and Fault Tolerant Controller for Boost Converter

Abstract: DC-DC power converters play an important role in energy conversion systems. Due to the age or unexpected conditions, sensors are prone to faults. The faulty sensors may cause wrong commands from the controller such that the actual voltage and current depart from the desired values. Therefore, there is a strong motivation to develop a fault-tolerant control scheme for power converters to improve their reliability. In this study, a boost converter is concentrated. It is demonstrated the existing linear fault estimator can effectively estimate sensor faults under open-loop operation condition but fail to estimate sensor faults in a closed-loop control circuit. A novel nonlinear fault estimator is addressed which can ensure an accurate estimate of both system states and sensor faults. A fault-tolerant design scheme using sensor signal compensation and nonlinear controller is proposed which can guarantee the closed-loop boost converter can track the desired voltage reference under sensor faulty conditions. The passivity of the tolerant control circuit is proved mathematically, and the effectiveness of the proposed algorithms are demonstrated by simulation studies.

Probabilistic wind power forecasting with an improved sparse-group Lasso-quantile regression neural network

Abstract: Probabilistic wind power forecasting is a critical technique in the decision-making process of the upcoming power grid with large penetration of environment-friendly energy generation. Conventional quantile regression based deep networks often suffer from quantile crossing and network redundancy. In this study, an improved non-crossing sparse-group-Lasso deep quantile regression model is presented for probabilistic wind power forecasting. Quantile crossing constraints and a sparse group Lasso algorithm are designed to pursue an interpretable and compact network. For illustration, an empirical case of a wind power dataset collected from a wind turbine SCADA system is used. Extensive comparisons validate the superiority of the presented model in terms of forecasting quality and computational efficiency.

Detecting Enclosed Board Channel of Data Acquisition System Using Probabilistic Neural Network with Null Matrix

Abstract: The board channel is a connection between a data acquisition system and the sensors of a plant. A flawed channel will bring poor-quality data or faulty data that may cause an incorrect strategy. In this paper, a data-driven approach is proposed to detect the status of the enclosed board channel based on an error time series obtained from multiple excitation signals and internal register values. The critical faulty data, contrary to the known healthy data, are constructed by using a null matrix with maximum projection and are labelled as training examples together with healthy data. Finally, the status of the enclosed board channel is validated by a well-trained probabilistic neural network. The experimental results demonstrate the effectiveness of the proposed method.

Spatial Optimization in Spatio-temporal Motion Planning

Abstract: Motion Planning is one of the key modules in autonomous driving systems to generate trajectories for self-driving vehicles. Spatio-temporal motion planners are often used to tackle complicated and dynamic driving scenarios. While effective in dealing with temporal changes in the environment, the existing methods are limited to optimizing a particular family of cost functions defined based on decoupled longitudinal and lateral terms. However, the planning objectives can only be explained using coupled terms in some cases, e.g. closeness to the reference path, lateral acceleration, and heading rate. The limitation arises from expressing such objectives as linear and quadratic terms suitable for optimization. This paper proposes an approach with theoretical proofs to approximate the upper bound of a given couple, nonlinear cost term with a set of uncoupled terms, allowing for converting the planning optimization problem into a linear quadratic optimization. The effectiveness of the proposed approach is shown through a series of simulated scenarios. The proposed approach results in smoother and steadier trajectories in the spatial plane.