To address the performance degradation of classical indirect field-oriented control of the induction motor caused by external disturbances, such as load torque variations and parameter uncertainties, this article proposes a continuous fixed-time backstepping control strategy based on a fixed-time disturbance observer using the super-twisting algorithm (FxTBC + FxTDO). The objective is to enhance the dynamic response and robustness of the induction motor drive system. First, a mathematical model of the induction motor incorporating total disturbance terms, including both external disturbances and parameter uncertainties, is established in the state-space framework. Next, a fixed-time disturbance observer (FxTDO) is designed to estimate these lumped disturbances. The estimated terms are then incorporated into the control law to compensate for their effects. Subsequently, a fixed-time backstepping controller (FxTBC) is developed. The fixed-time stability of the closed-loop system is rigorously established using Lyapunov theory. Finally, the effectiveness of the proposed approach is validated through MATLAB/Simulink simulations and compared with a super-twisting finite-time control scheme (FTC + FTDO). The results demonstrate that the proposed FxTBC–FxTDO strategy achieves a 20% reduction in settling time compared to the FTC–FTDO method, while providing superior disturbance rejection performance. Moreover, the chattering phenomenon is reduced by approximately 24% using the proposed approach.
Carton detection is a crucial task in intelligent warehouse systems, whose performance is primarily constrained by the limited storage and computing resources of warehouse edge devices. In this paper, a new model called BM-YOLO is proposed based on You Only Look Once 11 (YOLO11) to overcome these issues. First, a DimPool attention module is introduced into the backbone network, which integrates multidimensional collaborative attention with spatial pyramid pooling fast to enhance feature interactions and improve target discrimination. Second, a multiscale cooperative feature pyramid network is constructed in the neck network by combining the bidirectional feature pyramid network with multidimensional collaborative attention. This structure adaptively models cross-scale semantic relationships while reducing computational redundancy. Third, the original large detection head is removed to make the model lightweight without sacrificing detection accuracy. Experimental results in three public datasets validate the superiority of the BM-YOLO model compared with the baseline model YOLO11. In the live stacked carton dataset, BM-YOLO has achieved an increase of 33.2% in frames per second and an increase of 0.6% in precision of mAP@0.5. In the online stacked carton dataset, BM-YOLO achieves an improvement of 0.8% in precision along with an increase of 0.3% in both mAP@0.5 and mAP@0.5:0.95. In the logistics objects in context dataset, BM-YOLO achieves a 0.8% increase in precision and a 0.6% improvement in mAP@0.5, further validating the model’s generalization capability.
This paper investigates the trajectory tracking control of a powered parafoil in the presence of wind disturbances and actuator saturation. An anti-windup active disturbance rejection control method is developed on the basis of an eight-degree-of-freedom nonlinear model. For controller design and implementation, the system is organized into horizontal and vertical channels. The residual coupling, model uncertainty, and wind effects are not ignored; instead, they are incorporated into lumped disturbances and estimated online by an extended state observer. The key idea of the proposed method is to explicitly feed the mismatch between the commanded input and the saturated actuator output back into the observer update, so that actuator saturation is handled within the observer and controller loop rather than treated as a simple external limitation. A boundedness analysis is further provided to show that the closed-loop tracking errors are uniformly ultimately bounded under bounded disturbance derivatives and admissible gain selection. The proposed method is evaluated through same-platform simulations involving dynamic wind, horizontal saturation, thrust saturation, step disturbances, parameter variations, Monte Carlo tests, and sensitivity analysis. The results show that the proposed method is particularly effective in saturation-dominant cases and during disturbance transitions. Outdoor flight experiments further confirm its practical feasibility. Using the currently available flight data, the proposed anti-windup active disturbance rejection control reduces the maximum horizontal tracking error from 37.11 m to 8.11 m and decreases the altitude overshoot from 13.3% to about 8.0% compared with the deployed linear active disturbance rejection control baseline. These results demonstrate the practical value of saturation-aware observer correction for trajectory tracking of powered parafoils.
This article investigates fault-tolerant control and disturbance rejection for impulsive switched linear systems subject to periodic intermittent faults. First, a new periodic intermittent faults model is proposed, along with a Fourier series observer designed to accurately estimate faults and disturbances in the system. Next, a controller with disturbance rejection and fault-tolerant performance is designed based on these estimates. Stability is guaranteed under both arbitrary switching and average dwell-time strategies, and the correctness of this method has been demonstrated through numerical simulation and practical simulation.
In practical teleoperation engineering applications, operator negligence can easily lead to collisions of the remote robot in constrained environments. To address this, this research investigates the safety control and path following of teleoperated robotic systems in constrained task spaces, proposing a novel control framework that integrates an integral barrier Lyapunov function with a soft saturation function. The core of this framework is the construction of an adaptive learning controller based on integral barrier Lyapunov function. Unlike traditional approaches, the adopted integral barrier Lyapunov function here enforces constraints directly on the robot’s actual states instead of the tracking error. This ensures that the operation never violates the predefined spatial constraints. Meanwhile, by introducing a second-order filter and a soft saturation function on the remote side to generate a differentiable reference trajectory, it is ensured that any trajectory mapped from the master side remains strictly within the constrained interior. Furthermore, a radial basis function neural network is employed to approximate and offset modeling uncertainties in the system online, enhancing both robustness and tracking accuracy. Utilizing Lyapunov stability criteria, it is analytically demonstrated that both the local and remote controllers of the bilateral teleoperation system remain stable, with all state signals being semi-globally uniformly ultimately bounded. Finally, physical experiments are conducted to verify that the proposed method can ensure the remote robot moves within the prescribed constraints while providing force feedback. Comparative studies with other control methods further demonstrate its effective control performance.
This article addresses the trajectory tracking control problem for a class of nonlinear systems subject to nonparametric uncertainties. Traditional adaptive methods struggle with such uncertainties due to inadequate regression models, while conventional black-box neural networks suffer from poor generalization and physical inconsistency beyond the training distribution. To overcome these limitations, a physics-informed neural networks–based adaptive nonrecursive control framework is proposed. The core innovation lies in using the physics-informed neural networks mechanism to estimate the nonparametric uncertainty by embedding the system dynamics as a constraint in the loss function, ensuring physically consistent and highly generalizable uncertainty estimation even with sparse data. Furthermore, the physics-informed neural networks–based estimated uncertainty is directly fed into an adaptive nonrecursive controller to dynamically adjust the scaling gain, thereby achieving a composite estimation and control design that mitigates the analytical complexity growth and noise amplification in recursive methods. Rigorous stability analysis validates the uniformly bounded tracking error.
Accurate modeling of the turntable servo systems is crucial for high-precision motion control applications such as radar tracking and aerospace simulation. To improve modeling accuracy for nonlinear effects in turntable servo system under small-sample conditions, this paper proposes an attention-based deep kernel learning identification method. By introducing a multi-head attention mechanism into the feature extraction stage of deep kernel learning, the capability to extract nonlinear input features is strengthened, thereby improving the accuracy of nonlinear model identification. Since data on nonlinear factors in the turntable servo system are difficult to measure, an alternating compensation strategy between linear and nonlinear model errors is adopted on the basis of the attention-based deep kernel learning method. This enables the separate identification of linear and nonlinear models and ultimately improves the overall modeling accuracy of the turntable servo system. Simulation and experimental results demonstrate that, under small-sample conditions, the turntable servo system model prediction value obtained by the attention-based deep kernel learning identification method performs better than the deep neural network and deep kernel learning methods in terms of indicators such as mean squared error and quantitative fit.
Although some existing studies have addressed event-triggered iterative learning control–related problems, most of them focus on one-dimensional systems or investigate control systems with the aid of certain two-dimensional analytical methods. This article explores the problem of event-triggered adaptive iterative learning control design for two-dimensional linear discrete-time systems described by the Fornasini–Marchesini model. First, an improved event-triggering condition that only requires the system state and tracking error is devised. Second, by introducing the event-triggered control into the iterative domain, an adaptive parameter law is designed. Then, a novel two-dimensional event-triggered adaptive iterative learning control scheme is developed, which can effectively reduce the updating number of control inputs, alleviate the computational burden on the controller, and save network communication resources. A rigorous theoretical analysis is conducted on the developed two-dimensional event-triggered adaptive iterative learning control scheme using a composite energy function, and the results show that the developed scheme can ensure the convergence of tracking errors. Finally, a simulation example is given to demonstrate the effectiveness of the proposed two-dimensional event-triggered adaptive iterative learning control scheme.
The admissibility analysis and synthesis problems of nonlinear descriptor systems through extended T-S fuzzy singular systems with different derivative matrices are investigated. By designing the fuzzy Lyapunov function framework and introducing relaxation matrices, a novel and relaxed sufficient condition is presented to guarantee the admissibility of these systems without requiring bounds on the derivative of the membership function. Moreover, by establishing an admissibility equivalence between the original system and its dual system, a design theorem for the fuzzy parallel distribution compensation controller of fuzzy singular systems is proposed using strict linear matrix inequalities. This method is not only easy to implement, but also applicable to a wide range of similar problems. The effectiveness of the main results presented in this paper is illustrated through three illustrative examples.
In this paper, a sliding mode controller with a fractional-order integral surface is designed for a nonlinear quadrotor system affected by external disturbances. The finite-time stability of the suggested fractional-order integral sliding mode control idea is proved using the Lyapunov theorem. Then, effective parameters of the controller are identified using a multi-objective gray wolf optimization algorithm with regard to simultaneously minimizing system errors and control efforts. Finally, the obtained results are simulated to clarify the capability of the proposed approach to handle the dynamical nonlinearities and external disturbances.
Model order reduction for discrete-time systems is often required to preserve accuracy over a prescribed frequency interval rather than across the entire frequency range. It is observed that a few existing methods based on frequency-constrained Gramians yield significant approximation errors for discrete-time systems owing to the eigenvalue imbalance in some of the intermediate matrices. Therefore, in this paper, a novel model reduction technique is proposed for discrete-time systems by constructing new frequency-constrained Gramians. A novel set of pseudo-input and output matrices is formed that precisely approximates the higher-order system to a lower-order system within the specified frequency interval. The proposed method guarantees a stable reduced model for a given stable system and provides an a priori error bound for the desired frequency interval. The simulation results of numerical examples illustrate the effectiveness of the proposed technique.
Passivity is often useful for systems with an origin equilibrium. However, many practical systems have multiple equilibria or no equilibrium. This paper explores fixed-time incremental passivity for switched nonlinear systems using multiple incremental storage functions and applies this concept to address the fixed-time full-information output regulation problem. First, fixed-time incremental passivity is defined, allowing storage functions to increase at each switching time, ensuring that each active subsystem demonstrates fixed-time incremental passivity. Second, a more general switching law based on state information is provided to achieve fixed-time incremental passivity. This switching law employs continuous functions instead of storage functions, thereby offering greater design flexibility. Third, the developed fixed-time incremental passivity theory is utilized for fixed-time output regulation, even when individual subsystems lack fixed-time incremental passivity. Finally, the effectiveness of this control approach is illustrated through two examples.
Temperature prediction of traction motors plays a crucial role in the health assessment, safe operation, and early-warning maintenance of the traction motors for high-speed trains. In practical operation, traction motor temperature data are affected by varying operating conditions and usually exhibit non-stationary fluctuations, multi-scale characteristics, and coupled spatiotemporal dependencies. Aiming at the problems of insufficient feature extraction depth in traditional time-series prediction models and the failure of prediction results to meet engineering requirements under complex operating conditions, this paper proposes a traction motor temperature prediction framework integrating variational mode decomposition, a stacked bidirectional gated recurrent attention unit, a multi-channel temporal convolutional network, and cross-attention; the framework is termed variational mode decomposition-MTGAC. The model is designed to extract and fuse multi-scale spatiotemporal features in data, thereby achieving accurate prediction of traction motor temperature. First, variational mode decomposition is used to perform multi-scale decomposition on the original monitoring data, which eliminates noise and redundant components and improves data quality. Second, a temporal feature extraction module based on a stacked bidirectional gated recurrent attention unit is designed to capture the long-term and short-term dependencies in the data and extract the features of temperature changes at different timescales. Meanwhile, a spatial feature extraction module based on a stacked multi-channel temporal convolutional network is constructed; by leveraging the parallel computing advantage of multi-channel convolution kernels, this module mines the spatial features among temperature changes from multiple dimensions. Finally, a cross-attention mechanism is adopted to deeply fuse spatiotemporal features and dynamically adjust weights to highlight key information, thus enhancing the model’s feature representation capability. A dataset is constructed using the actual operation data of CR300BF electric multiple units, and comparative experimental tests are conducted under multiple prediction scenarios. The experimental results show that under the two scenarios with prediction output steps of 45 and 90 min, compared with Informer, Bidirectional Gated Recurrent Unit, Convolutional Neural Network, and Transformer models, the variational mode decomposition-MTGAC model achieves an increase of 1.10%–3.95% in R 2 and reductions of 24.90%–73.99% and 42.34%–68.41% in mean squared error and mean absolute error, respectively. In different prediction scenarios, the temperature prediction curve of the proposed model has a higher fitting degree with the actual temperature curve. It not only improves the prediction accuracy and model generalization ability but also reduces the prediction error, providing technical support for traction motor temperature prediction and early-warning applications in high-speed trains.
The high-frequency (HF) signal injection method is widely used in sensorless control of permanent magnet synchronous motors (PMSMs). Conventional strategies typically integrate a band-pass filter (BPF) with a phase-locked loop (PLL) to extract HF current responses and estimate rotor position. However, this approach introduces significant phase delay, restricts system bandwidth, and exhibits poor robustness against disturbances. To overcome these limitations, a strategy that integrates a second-order generalized integrator (SGI) with a linear extended state observer (LESO) was proposed. The SGI enables rapid demodulation of the HF current response with negligible phase lag, while the third-order LESO provides strong disturbance rejection by estimating and compensating for total system uncertainties in real time. The validity of the proposed sensorless control strategy has been demonstrated through simulations and experimental tests. Simulation and experimental results show that the SGI-LESO significantly reduces position-estimation error, improves disturbance rejection, and ensures smoother start and dynamic speed tracking.
To achieve trajectory tracking control of a robotic manipulator system, it is crucial to establish precise kinematic and dynamic models. Considering the complex characteristics of robotic manipulator, a dynamic modeling method based on virtual decomposition control is proposed. Considering the interaction force between the end effector of the robotic manipulator and the environment, an adaptive controller-nonlinear extended state observer combination control is designed through virtual decomposition control theory. This method virtually decomposes the complex robotic manipulator system into several subsystems, achieving adaptive controller-nonlinear extended state observer combination control for each subsystem. By introducing the concept of virtual power flow into the controller and designing Lyapunov functions based on nonlinear extended state observer error dynamics, the stability of this redundant system is proven.
To address the issues of low convergence accuracy and deteriorated output performance caused by initial state errors in linear time-invariant discrete-time systems, traditional iterative learning control methods typically treat initial errors as passively tolerated conditions. The lack of an active correction mechanism significantly limits the dynamic performance of the system. To overcome this limitation, this paper proposes a proportional-–integral-–derivative–type iterative learning control algorithm based on the normalization principle, incorporating a dynamic compensation term constructed via an inverse tangent function. First, by analyzing the impact mechanism of initial state errors on system output, a gradually decaying dynamic compensation term based on the arctangent function is designed under the normalization principle. This term is integrated into the control gain to enable active learning and smooth correction of initial errors. Second, the convergence condition of the proposed algorithm is derived, and the tuning guidelines for key parameters are provided. Theoretical analysis demonstrates that the system output error converges monotonically in the iteration domain. Finally, numerical simulations verify the effectiveness of the proposed algorithm. The results show that, under identical system conditions, the proposed method reduces the system output error by more than 80% compared to existing typical approaches, significantly improving both convergence speed and steady-state accuracy.
This paper addresses the time-varying formation (TVF) control problem for linear multi-agent systems (MASs) with nonlinear dynamics, actuator bias and loss-of-effectiveness faults, and switching communication topologies. First, a novel distributed observer is developed to estimate follower states via neighboring interactions. Then, a controller is designed to achieve the desired TVF tracking. By neural network (NN) approximation and Lyapunov stability theory, adaptive control gains are derived under a formation tracking feasibility condition. Finally, simulation results validate the derived theoretical results.
To address the challenges of insufficient abnormal samples, data imbalance, and poor generalization in industrial fault diagnosis, this study proposes a universal fault sample augmentation framework integrating adaptive signal reconstruction and multi-scale stochastic quantization. The framework comprises five core modules: adaptive signal decomposition, optimized cost function construction, sliding window segmentation, multi-scale stochastic quantization, and clustering threshold constraints. Multi-objective optimized variational mode decomposition (VMD) via Limited-memory Broyden–Fletcher–Goldfarb–Shanno with Bound constraints is employed to extract stable and physically interpretable components, reducing manual parameter dependence. A multi-scale sliding window mechanism combined with Gaussian-based stochastic quantization enhances sample diversity while preserving intrinsic signal characteristics, and a clustering-based constraint is introduced to filter pseudo-anomalies. Experimental results on multiple benchmark datasets demonstrate that the proposed method significantly improves diagnostic performance and achieves robust generalization under imbalanced conditions.
Regarding the stability control problem of uncertain Takagi-Sugeno (T-S) fuzzy descriptor systems, a non-fragile controller design method based on the collaborative scaling of multi-parameter uncertain terms is devised in this paper. First, through the joint scaling of system parameter uncertain terms Δ A and the boundaries of the derivative functions of membership functions, the system stability problem is converted to a problem that can be solved by Linear Matrix Inequalities (LMIs). Moreover, the admissibility of the system is rigorously verified. Second, relying on the analysis results, a non-fragile controller is designed, and a dynamic scaling mechanism for the gain uncertain term Δ K is introduced to mitigate the conservatism of the control system. Finally, simulation validation covers basic uncertain systems, complex scenarios with gain uncertainties, as well as practical systems such as inverted pendulum cart systems and biological population systems. Quantitative comparisons demonstrate that the proposed method ensures asymptotic stability of the system; meanwhile, it outperforms existing approaches in terms of the range of feasible regions, the tolerated limit robust uncertainty threshold, and dynamic response efficiency, providing new ideas for the robust control of non-linear engineering fuzzy systems.
Fault diagnosis in electric vehicle motor drive systems is essential to ensure reliable and safe operation. Especially in voltage source inverter-fed permanent magnet synchronous motor drives, switch faults need attention as they are the most fault-prone. Furthermore, the power switch faults, including open-circuit and short-circuit conditions, can significantly affect system performance. However, existing model- and signal-based fault diagnosis techniques often face limitations such as dependency on precise modelling and often give limited accuracy under dynamic operating conditions. On the contrary, existing data-driven approaches experience difficulties from feature redundancy and reduced generalisation in high-dimensional datasets. To address these challenges, this paper proposes an ensemble subspace k-nearest neighbour model for fault detection and classification of switch open-circuit and short-circuit faults. From a developed simulation model, the datasets under normal and multiple fault conditions (F0–F8), incorporating variables such as three-phase currents, are collected. Furthermore, the statistical features are extracted and subsequently reduced using principal component analysis to enhance feature representation and computational efficiency. The proposed model is evaluated using fivefold cross-validation and compared with other machine learning approaches, including bagged trees, quadratic support vector machine, boosted trees, and a neural network. The results reveal that the proposed subspace k-nearest neighbour model attains a testing classification accuracy of 98.49%, with an average precision of 97.5%, a recall of 98.3%, and an F1 score of 97.8%, surpassing the comparative models. These results denote that the proposed subspace k-nearest neighbour provides an effective and computationally efficient solution for reliable fault diagnosis in electric vehicle motor drive systems.
