International Journal of Adaptive Control and Signal Processing最新文献

This paper focuses on the parameter estimation issue of the nonlinear exponential autoregressive model with non-Gaussian noise. By exploiting the property that the Cauchy kernel correntropy is insensitive to non-Gaussian noise and the kernel bandwidth, a Cauchy kernel correntropy-based criterion function is presented to obtain robust estimation performance. Since the nonlinear exponential autoregressive model includes the coupling term of the linear and nonlinear parameters, the identification model is decomposed into two fictitious sub-models with a common parameter vector. A Cauchy kernel correntropy-based robust partially coupled recursive least squares stochastic gradient algorithm is proposed by coordinating the coupling term arising from the model decomposition. Compared with the existing recursive least squares and least mean squares algorithms, the proposed algorithm not only demonstrates robustness to non-Gaussian noise but also achieves higher estimation accuracy. The simulation examples exhibit the effectiveness of the proposed algorithm.

The cubature Kalman filter (CKF) is widely used for state estimation in nonlinear systems but struggles with noise uncertainties, non-Gaussian measurement noise, and outliers. This paper introduces a novel Q-learning-based adaptive Student's $��t�$-distribution maximum correntropy CKF (QL-STMCCKF) to address these issues. By integrating reinforcement learning and maximum correntropy criteria with the CKF, the method improves robustness. The Student's $��t�$-distribution kernel handles heavy-tailed non-Gaussian noise, while Q-learning adaptively updates process and measurement noise covariances, enhancing estimation performance. A maximum likelihood estimation (MLE)-based variant, MLE-STMCCKF, is also developed for comparison. The algorithms are tested on a benchmark aircraft tracking problem and validated with offline hardware data from a grid-connected 3-phase voltage source inverter. Comparative analysis highlights the superior performance of the QL-STMCCKF in challenging conditions.

In this paper, a non-linear adaptive direct current controller based on periodic estimation is proposed to significantly minimize the periodic torque ripples for the permanent-magnet synchronous motor system in the discrete-time setting. The discrete-time adaptive controller is designed to track the desired motor currents of motor and provide periodic disturbance rejection by constructing the estimation algorithm based on the digital processing of unknown periodic signals in each consecutive motor period, which requires only knowledge of periodicity. The asymptotic stability of the closed-loop system dynamics is proven through the convergence of current errors to zero as the period of the motor shaft approaches infinity. To test the robustness against periodic uncertainties and disturbances varying with the rotor displacement and to demonstrate the achievement of the desired steady-state response for stator current quality, detailed simulation experiments are carried out with comparisons to the classical deadbeat controller.

The accurate prediction of wind power is crucial for enhancing the integration of wind energy into the grid. Wind power prediction models require high-dimensional input to ensure reliable results. However, challenges arise in obtaining wind power data due to instrument failures. To achieve accurate forecasting, it is essential to fill in missing data and accurately interpret dynamic properties. This study addresses issues such as missing data from sensor failures by introducing the ExZe-DenseLSTM model for precise wind power predictions. The model collects important variables like variance, standard deviation, kurtosis, skewness, and correlation and uses the K-Nearest Neighbors (KNN) technique for data imputation. For better performance, the model mixes DenseNet with Long Short-Term Memory (LSTM), and feature selection is carried out using the Improved Wrapper Approach. To increase the wind power forecasting model's training efficiency and prediction accuracy across a range of learning rates, the Extended Zebra (ExZe) algorithm is used to optimize the model's loss function. The proposed method outperforms existing prediction techniques, producing better outcomes with an MAE of 0.180, MAPE of 107.21, and MSE of 0.094, respectively.

This paper presents a novel neural adaptive sliding mode control (ASMC) for robust quadrotor trajectory tracking under severe nonlinearities, measurement noise, and external disturbances. The proposed strategy integrates a neural feedback-error-learning (FEL) framework that adaptively tunes both the sliding surface and reaching law gains online via a dual-cost optimization scheme, thereby simultaneously reducing tracking errors and control effort. Two key contributions form the cornerstone of this work. First, the neural FEL framework enables the online optimization of control gains by employing distinct cost functions that separately address tracking performance and control effort. Second, the integration of a flexible sigmoid activation function in the ANN's output layer guarantees strictly positive control gains, which enhances system stability. Simulation studies demonstrate that, compared with a similar ANN-based ASMC approach, the proposed controller achieves approximately a 26% reduction in maximum overshoot, nearly a 49% decrease in settling time, and roughly a 64% reduction in steady-state error. In addition, significant improvements in tracking performance are observed, as evidenced by reduced root mean square error (RMSE) and mean absolute error (MAE) while maintaining comparable control effort. These quantitative enhancements underscore the potential of combining adaptive neural tuning with sliding mode control for high-performance and reliable quadrotor operation in complex, uncertain environments.

This paper investigates the problem of fuzzy adaptive predefined-time full-state error constraints for permanent magnet synchronous motor systems. This study employs fuzzy logic systems (FLSs) to handle unknown nonlinear dynamic functions. A state observer based on fuzzy logic systems is utilized to estimate unmeasurable states. By utilizing a general potential Lyapunov function, predefined-time stability theory, combined with the dynamic surface control (DSC) technique and backstepping recursive technique, a predefined-time full-state error constraints control strategy is proposed. By applying predefined-time Lyapunov stability theory, it is proven that all signals in the closed-loop systems remain bounded and the tracking error converges within a predefined time. Finally, the effectiveness and feasibility of the proposed method are verified through simulation results.

This paper discusses the adaptive control problem for a class of stochastic nonlinear systems with uncertain nonlinear functions and uncertain measurement functions. First, a series of neural network functions are used to estimate the unknown nonlinear terms. Then reasonable assumptions about the unknown powers $��{�q�}_{�i�}�$'s are raised based on the notion of the homogeneity with monotone degrees. By recursively constructing twice continuous differential Lyapunov functions, the adaptive feedback controller is designed to deal with the unknown coefficients. Based on the stochastic stability theorem, it is proved that all signals in the closed-loop system are bounded in probability. Furthermore, the effectiveness of the proposed control approach is verified by a practical example and a numerical simulation.

A five-dimensional hyperchaotic system is proposed based on a three-dimensional autonomous chaotic system by introducing two new variables. Based on Lyapunov stability theory and passivity theory, the investigation looks into passivity-based adaptive chaotic synchronization. Through theoretical analysis and numerical simulation, the dissipation, phase diagrams, Lyapunov exponents, and bifurcation diagrams of the system were studied in detail to analyze its dynamic characteristics. The analog circuit simulation of the new five-dimensional system is carried out by Multisim. The simulation results are verified by comparing them with the numerical simulation results in MATLAB. The adaptive control method combines passive and adaptive control theories to provide a new perspective and solution for the synchronization control of chaotic systems. The parameter estimation and updating laws not only improve the flexibility of synchronization control but also provide strong support for subsequent experimental studies and practical applications. The effectiveness of the proposed method is verified by numerical simulation, which proves that the method performs well in coping with the synchronization problem of complex hyperchaotic systems with uncertain parameters.

Accurate and reliable wind power forecasting plays a crucial role in the utilization of wind energy. However, the complexity of the nonlinear process of converting wind energy into power alters the statistical distributions of errors (known as concept drift), making the accomplishment of this task greatly challenging. For this purpose, we devise an innovative approach for predicting wind power based on the “decomposition-prediction-ensemble” framework. First, the raw wind power data is broken down into multiple intrinsic mode functions (IMFs) via improved variational mode decomposition, and the dimensionality of these IMFs is reduced by adopting kernel principal component analysis, thus significantly simplifying the intricacies of the multidimensional IMF data. Then, considering the asymmetric characteristic of modeling error, a probability density function control- based temporal convolutional network is developed for each subseries, where the modeling error PDF is controlled to close to an ideal Gaussian distribution, so that the prediction model parameters are adjusted. Finally, the multiple subseries forecasting models are integrated by a PDF-based differential evolution ensemble strategy. And the convergence of ensemble strategy is analyzed from a mathematical point of view. As a result, the proposed method can break through the limitation of symmetric loss capturing merely the second moment information, alleviating distribution shift of error to make an unbiased estimate. Real-world wind farm data are utilized for empirical analysis, and the simulation outcomes verify the high accuracy and generalization ability of the proposed model.

This paper presents a novel dynamic predictor-based event-triggered control strategy to investigate the mean-square exponential stabilization for a class of stochastic systems with time-varying output delay. First, a dynamic prediction scheme is introduced to deal with the problem of time delay, where the segmentation precision is no longer a constant, but a dynamically changing value. Second, an event-triggering mechanism with time regularization is proposed to avoid the Zeno phenomenon and guarantee the exponential convergence of the segmentation precision in the prediction scheme. Then, a predictor-based event-triggered control is designed to achieve the mean-square exponential stabilization of stochastic systems. Finally, a single-link robot model is given to show the effectiveness of the obtained results.