共找到 20 条结果
暂无摘要（点击查看详情）
In this paper, observer-based state feedback stabilization problem for a class of strict-feedback nonlinear systems with periodically event-triggered sampled-output-data is investigated. The output of the nonlinear system is sampled periodically but is monitored and determined to be transferred or not by a deliberately designed event-triggered mechanism. With the transmitted sampled-output-data, the unmeasurable states are estimated by a delicately constructed reduced-order observer and the state feedback stabilizer, elaborately designed by utilizing the homogeneous domination approach, is implemented to make the closed-loop nonlinear system asymptotically stable. The proposed method not only introduces a periodically event-triggered mechanism to determine the time-varying sampling instants, instead of just pointing out their existence, to update the designed observer and controller, but also alleviates the conservatism in stability analysis of closed-loop nonlinear system, which does not require the energy function to monotonically decrease, only to be bounded at inter-sampling intervals. Lastly, simulations are provided to demonstrate the efficacy of the proposed approach.
This paper investigates the fault estimation and fault-tolerant consensus tracking control of multi-agent systems (MASs) with leader of unknown input. Multi-reduced-order intermediate observers are constructed to estimate the state and unknown inputs of leader, along with the follower states, actuator and sensor faults. Compared with the existing results, the centralized and the distributed output estimation errors are added to the observers to improve the estimation performance. Utilizing estimated information, the distributed consensus protocol is constructed, which can compensate for the faults and leader's unknown input, and achieve consensus tracking of MASs. The parameter solution method is proposed, whose computational complexity is equivalent to that of a single agent. The root-mean-square (RMS) performance is introduced, which can reduce the impact of disturbances, and there is no need for the classical assumption of zero initial conditions. Finally, the feasibility of the method is verified through two examples.
To address the challenges posed by the high dimensionality and strong nonlinearity of complex industrial process data, as well as the inherent defects in feature learning and hyperparameter dependency of the traditional long short-term memory network (LSTM), this paper proposes a novel hierarchical fault diagnosis method termed entropy space quantum behaved dung beetle optimized LSTM (ES-QLSTM). The proposed method employs kernel entropy component analysis (KECA) to extract features from raw data in the entropy space (ES). Benefiting from the strong nonlinear mapping capability and high-order information retention property of the KECA, the proposed method effectively preserves critical high-order statistical information that cannot be captured by conventional methods, which compensates for the insufficient feature mining ability of the LSTM. Then, this proposed method integrates entropy space feature extraction with a quantum behaved mechanism into the dung beetle optimization algorithm, constructing a systematic synergistic framework for entropy space feature dimensionality reduction and intelligent hyperparameter optimization. This framework solves two key shortcomings of the LSTM simultaneously, weak feature extraction from high-dimensional nonlinear data and unstable performance caused by empirical hyperparameter settings. Subsequently, a quantum behaved mechanism is incorporated into the dung beetle optimization (DBO) algorithm, resulting in the quantum behaved dung beetle optimization (QDBO) algorithm, which is utilized to optimize the key hyperparameters of the LSTM network. The ultimately constructed ES-QLSTM method enhances diagnostic performance through the synergistic integration of entropy-space feature purification and intelligent hyperparameter optimization. Experimental validation on the Tennessee Eastman Process (TEP) and grid-connected photovoltaic system (GPVS) datasets demonstrates that the proposed method outperforms traditional models and improved models in diagnosing complex faults, with the average diagnostic accuracy reaching 93.70% and 83.91% on the two datasets, respectively.
In recent years, compound noise, including additive noise and multiplicative noise, has inevitably affected stochastic multi-agent systems (SMASs). These noises introduce unpredictable oscillations that severely degrade system stability. This influence becomes particularly critical in containment control problems where multiple leaders exhibit dynamic behaviors. Therefore, this article addresses this challenge by introducing a novel model that incorporates compound noises and the multiple dynamic leaders can evolve dynamically through interactions with their neighboring leaders. In order to reduce the effect of compound noises, based on stochastic approximation (SA) technique, a novel containment control protocol is designed. Since the introduction of multiplicative noise causes existing error analysis methods to fail for our proposed SMASs with multiple dynamic leaders, a novel semi-decomposition technique is proposed to achieve containment control in a compound noisy environment, where followers converge to the convex hull spanned by dynamic leaders. Additionally, the containment control problem with multiple static leaders scene can be viewed as a special case of our underlying system. Different from existing results for control gain conditions (∫0∞σ(t)dt=∞ and ∫0∞σ2(t)dt<∞), a weaker condition of ∫0∞στ(t)dt<∞ is adopted, where τ=min{2,ρ}>1 rather than 2. Then, three numerical simulations are proposed to illustrate the feasibility of our results. By selecting the control gains as ξ(t)=1/(t+1)0.8 (leaders' gain) and σ(t)=1/(t+1)0.6 (followers' gain), that is, the convergence speed of the leaders is faster than that of the followers, the containment control for SMASs with multiple dynamic leaders can be achieved. Additionally, the noise intensities of multiplicative noise have a greater impact on the convergence than those of additive noise.
This paper addresses the prescribed-time cooperative guidance problem for striking a target under time-varying velocity, leveraging deep neural networks to enhance prediction accuracy. Unlike existing results, the proposed guidance law guarantees the stability of the guidance error system even under switching topologies. First, a 3-D vector-based cooperative guidance model is established, and the cooperative guidance objective is formulated. To achieve precise time-to-go estimation under time-varying velocity, a high-precision prediction algorithm based on deep neural networks is developed. Building on this, a practical prescribed-time cooperative guidance law accounting for time-varying velocity is designed, with rigorous stability analysis provided for the guidance error system under switching topologies. Furthermore, based on guidance trajectory analysis, 2-D results in both horizontal and vertical planes are provided. Finally, the proposed method is validated through numerical simulations and equivalent physical experiments.
The effective fault diagnosis of the air handling unit (AHU) is critical for maintaining indoor environmental quality and reducing building energy consumption. However, existing fault diagnosis methods often fail to fully capture spatio-temporal features from AHU's operational data and struggle to maintain classification performance when different fault modes present distinct feature distributions. To address these challenges, this paper proposes a spatio-temporal feature extraction with an enhanced ensemble adaptive classification strategy, named STFE-SDTEL. Firstly, a spatio-temporal parallel transformer architecture is designed to simultaneously capture temporal dependencies and spatial features, where a parallel gated convolutional neural network branch is incorporated to enhance spatial feature extraction and suppress redundant information, thereby overcoming the limitations of traditional transformer in spatial feature modeling. In addition, to strengthen the interaction and fusion of temporal and spatial features, a bidirectional cross-attention mechanism is introduced, enabling the model to learn more discriminative spatio-temporal representations. Finally, to improve classification adaptability to heterogeneous feature distributions across different fault modes, an enhanced ensemble adaptive classification strategy is developed, which dynamically combines multiple soft decision tree outputs using adaptive weights. Experimental results on the ASHRAE RP-1312 dataset show that STFE-SDTEL achieves an overall diagnostic accuracy of 99.68%. Comparative results with several mainstream baseline models further demonstrate the superior diagnostic performance of the proposed method.
This paper presents a robust fixed-time sliding-mode control strategy for an electro-hydraulic active suspension (EHAS) system with nonlinear actuator dynamics and external road disturbances. Due to the hydraulic subsystem, the considered quarter-car model exhibits a relative degree of four with respect to the sprung-mass displacement, which makes fast and robust vibration attenuation particularly challenging. The proposed controller ensures fixed-time convergence of the sliding variable, independently of the initial conditions, while explicitly accounting for model uncertainties and disturbances. The control performance is evaluated through numerical simulations under several representative road profiles, including sinusoidal excitation, speed hump, pothole, and broadband chirp inputs. The simulation study incorporates realistic operating conditions, including actuator saturation, nonlinear hydraulic dynamics and measurement noise. Ride comfort is assessed using ISO 2631-1 criteria, together with road-holding, suspension travel, tracking accuracy, and control smoothness metrics. Comparative results show improved transient performance and ride comfort with respect to classical sliding-mode, backstepping, and PID controllers, while maintaining satisfactory suspension safety and control effort.
For the challenge of cooperative optimization of transient performance, energy consumption, and communication resources in multi-agent systems, this paper proposes an event-triggered prescribed time optimal consensus control scheme within the framework of deep reinforcement learning-based optimal backstepping. Firstly, a distributed event-triggered communication mechanism is proposed by designing the output sampling event-driven function with bandwidth sensing characteristics to realize the elastic adjustment of the communication load from the topological dimension. Subsequently, to find the optimal control solution for the co-optimization of stability and energy consumption, an optimal consensus control protocol is constructed based on the actor-critic neural network iterative learning algorithm and Bellman optimality theory. Furthermore, by introducing a time-varying gain scaling function, the Hamilton-Jacobi-Bellman equation analytical framework with explicit time-constrained characteristics is reformulated and an optimal consensus controller with a strict prescribed time convergence guarantee is derived that balances transient performance, energy consumption, and communication resources while attaining multi-agent system optimal consensus. Finally, the effectiveness of the proposed scheme is validated through comparative numerical simulations and a model simulation of a multi-agent electromechanical system.
This paper addresses the distributed multi-objective optimization problem for discrete-time heterogeneous multi-agent systems via potential games. Potential game-based methods are widely employed in distributed optimization for multi-agent systems, as they enable the decoupling of solution processes and ensure convergence to the desirable equilibrium. While potential game theory is effective for single-objective cases, its multi-objective extension lacks a systematic framework, is limited by strict assumptions, and exhibits poor explainability. To overcome these limitations, a novel semi-tensor product (STP)-based framework is proposed, which is one of the powerful tools for the research of finite potential games. The main contributions are (1) formulating a novel game model-finite multi-objective networked potential games (MONPGs)-for heterogeneous interactions, with an STP-based algebraic condition enabling local information-based payoff design; (2) designing a strategy learning algorithm that guarantees the convergence to a Pareto equilibrium and is universally applicable to arbitrary real-valued payoff vectors, significantly enhancing generality compared to prior works; and (3) deriving a sufficient condition expressed as a linear matrix equation for solving the distributed optimization problem. This work extends potential game-based methods to multi-objective and heterogeneous interaction settings, enhancing interpretability and solving a class of problems previously intractable for existing potential game-based methods.
This article addresses the global exact prescribed-time output feedback stabilization problem for a class of nonlinear systems subject to measurement uncertainty and unknown nonlinearities. Unlike many prescribed-time control results, where the theoretical analysis does not exclude the possibility of convergence before the prescribed time, the exact prescribed-time convergence considered in this paper further characterizes the non-premature convergence: nonzero closed-loop trajectories approach the origin at the prescribed time in the limiting sense and do not vanish before that time. To achieve this guarantee, the proposed approach employs the solutions of the parametric Lyapunov equations (PLEs) as the central analytical tool and fully exploits their properties, in conjunction with time-varying Lyapunov-like functions, to construct a linear observer-based output feedback control strategy. A salient feature of the proposed scheme is that it relies exclusively on linear time-varying gains. Finally, two numerical case studies, including comparisons with existing methods, are provided to demonstrate the effectiveness of the proposed approach.
Complex traffic scenes with dynamic occlusions create hidden blind-spot hazards that challenge the reliability of autonomous driving decision control. This study aims to improve safety under such uncertainty by introducing a meta-reasoning-driven dual-level decision framework. The framework integrates object-level and meta-level processing to assess situational uncertainty and adapt decision responses accordingly, and incorporates a hippocampus-amygdala-inspired importance-sampling mechanism to enhance learning from error-prone cases. Systematic experiments show that the proposed method reduces collision risk and yields more reliable and efficient decision behavior compared with conventional approaches. Across multiple experimental settings, the proposed framework improves the success rate by approximately 2-18% and increases the driving index by about 4-23% relative to baseline methods. These results demonstrate that meta-reasoning can effectively strengthen safety-critical decision control in uncertain environments and provide a feasible direction for incorporating biologically inspired mechanisms into autonomous driving systems.
To address the challenge of balancing finite-time convergence and chattering suppression in sliding mode control for permanent magnet synchronous motors, this paper proposes a composite control scheme based on an adaptive three-stage reaching law (ATSRL) and an enhanced super-twisting sliding mode observer (ESTSMO). First, according to the distinct characteristics of system state, the ATSRL incorporates a clamped exponential term, a state-dependent adaptive term, and a nonlinear power term. It adaptively adjusts the reaching speed by system state and reduces the switching gain near the sliding surface, enabling fast finite-time convergence while effectively suppressing chattering. Second, to mitigate the deterioration of control performance caused by inertia mismatch, an extended sliding mode observer is employed for offline identification of system inertia, and then the ESTSMO is devised to estimate lumped disturbance online as a feedforward compensation term, which is directly embedded into the ATSRL-based control law. Interestingly, the offline-then-online framework provides accurate offline inertia identification via the ESMO, which is incorporated into the ESTSMO to significantly reduce disturbance estimation bias caused by inertia mismatch. Finally, simulation and experimental results validate the advantages of the proposed scheme.
In this article, an advanced nonlinear control strategy is proposed to address the trajectory tracking problem of a quadrotor under external disturbances. The controller is formulated as an Adaptive Fractional-Order Sliding Mode Controller (AFOSMC), which employs a super-twisting reaching law to achieve robust tracking, while an adaptive barrier function is incorporated to alleviate the effects of control signal saturation. In addition, the conventional discontinuous signum function is replaced with a smooth saturation function to effectively mitigate the chattering phenomenon. To further strengthen the performance of AFOSMC, ant colony optimization is used for tuning the controller gain parameters. The quadrotor dynamics are derived using the Euler-Lagrange formulation, providing a rigorous basis for controller design. Lyapunov-based analysis is conducted to verify system stability, ensuring reliable operation under varying conditions. Additionally, stochastic noise is incorporated into the quadrotor model to investigate the proposed controller's robustness. The proposed strategy is evaluated via numerical simulations, while its practical applicability is further assessed using controller-in-the-loop experiments. The obtained findings indicate that AFOSMC provides improved tracking precision while maintaining robustness and smooth control action.
Generally, there are always many unknown disturbances in the aero-engine, such as the complex aerodynamic load force, the bandwidth and computation resource limitations. If these uncertainties are not effectively addressed, they will seriously affect the safety and performance of the hydraulic system. In this paper, the radial-basis-function (RBF) neural network (NN) combined with the event-triggered technique is used to address these unknowns, in which NN weights update only at necessary instants. This permits a greater number of NN nodes to be employed, augmenting the estimation accuracy of RBF-NN estimators without overconsuming limited computational resources. A new motion control scheme with an event-triggered communication protocol and the designed estimator is integrated to ensure the high performance and low occupation of bandwidth resources for hydraulic actuators. Using the proposed controller for hydraulic systems and analyzing from the viewpoint of impulsive dynamical systems, it is concluded that all closed-loop signals are ultimately bounded and the system output can accurately track the reference trajectory. Likewise, Zeno behavior no longer occurs. Finally, two comprehensive experiments, i.e., the unknown mass load and aerodynamic load force, demonstrate the remarkable validity and merits of the proposed control scheme in the aero-engine Hardware-in-the-Loop (HIL) networked control platform with hydraulic actuators.
Spectral coherence theory is of practical importance for bearing fault diagnosis. However, when fault-related information is distributed across multiple spectral bands, existing methods often cannot effectively identify and integrate these informative components, which limits diagnostic performance. To address this issue, an optimal weighted envelope spectrum (OWES) is proposed. First, the expected signal to expected noise (ESEN) is constructed. By introducing local background correction and adaptive harmonic position identification, ESEN enhances the robustness of fault feature extraction to frequency deviation, amplitude variation, and spectral background fluctuation. Then, informative spectral regions are identified by thresholding the ESEN distribution along the spectral frequency axis and are merged into candidate frequency bands. A combinatorial optimization strategy is further introduced to select the candidate band subset that yields the most prominent integrated fault feature, thereby improving fault feature extraction when informative components are distributed across multiple spectral bands. Finally, the selected bands are weighted according to their information contributions and integrated to construct the OWES. The proposed method is validated using one simulated signal and three bearing experimental datasets. Compared with several advanced methods, OWES achieves clearer identification of fault characteristic frequencies and harmonics. Quantitative comparisons based on ESEN and kurtosis further demonstrate the superior performance of the proposed OWES method.
Induction motors (IMs) are widely used in energy conversion and industrial drive systems. One of the critical faults in IMs is the broken rotor bar (BRB), which must be promptly identified due to its potential for abrupt damage. Several monitoring techniques have been proposed for the detection process, including motor current signature analysis (MCSA) and model-based fault detection (MBFD) but the challenges of supply unbalance and harmonic conditions persist as a gap. This paper presents a BRB detection framework based on Conservative Power Theory (CPT), employing time-domain power components as physically interpretable features. In contrast to spectral-based approaches, the proposed method extracts statistical features from the CPT power quantities, naturally accommodating nonsinusoidal and unbalanced supply conditions. The fault identification is evaluated using the k-nearest neighbors, support vector machines, random forest, and multilayer perceptron classifiers. Experimental validation was conducted on three induction motors operating at 50 Hz and 60 Hz, under multiple loads and voltage unbalance factors up to 5.77%. The results demonstrate performance metrics of more than 98% in all cases, with maximum accuracy and an F1-score of 99.77% in binary classification. The treatment of incipient fault severity is addressed by assessing the likelihood of a motor being healthy or faulty, allowing the severity and progression analysis of the faults.
In this study, a robust dual-loop hybrid Fractional Order Tilt Integral Derivative (FOTID)-Proportional Derivative (PD) controller is developed to track both step and ramp signals. Here, the inner loop is an integer-order PD controller that shifts the process pole away from the origin, and the outer loop uses an FOTID controller. The proposed control scheme consists of only three tunable parameters, determined through an analytical design procedure supplemented by a simple one-dimensional graphical/numerical selection (via Ms-Kp characteristics), thereby reducing the overall design intricacy. The controllers are tuned using maximum sensitivity (Ms) and stability margins to provide good robustness in the closed-loop response. Using these design criteria, analytical solutions of the controller parameters are also obtained in terms for plant parameters. Case studies on a heat exchanger, a higher-order integrating process, and a two-area time-delayed cyber-physical power system have been undertaken to validate the effectiveness and reliability of the proposed control mechanism. The robustness of the suggested controller is examined by introducing a 20% perturbation to the plant parameters. Moreover, the simulated results are evaluated with respect to different errors. It is worth noting that for stable first-order plants, the proposed FOTID-PD framework reduces to an FOTI controller, since the inner PD loop is not necessary for plant stabilization. This simplified structure is used for experimental validation. Lastly, the proposed methodology is experimentally validated on a two-tank level loop.
The stochastic nature of time delays in networked control systems poses significant challenges for controller synthesis and the corresponding analyses, leading to conservative designs and degraded performance. Existing approaches approximate stochastic delays by fixed, worst-case values, which limits their ability to describe the behavior of distributed control implementations, including distributed computational architectures, multi-core processing, and network communication between different computational nodes. Thus, with regard to this context, this work proposes a novel modeling framework for linear multiple-input multiple-output networked control systems that represents stochastic sampling instants and delays through a stochastic linear time-varying state-space model. Based on this model, a predictor-based compensator derived from the filtered Smith predictor is developed to mitigate the effects of stochastic time delays. Using a cooperative adaptive cruise control benchmark, the proposed compensator is compared with a baseline fixed-delay predictor in terms of performance degradation, with all metrics normalized relative to the delay-free case. Numerical results indicate that the proposed method achieves a 55% reduction in the worst-case tracking error energy with respect to the baseline controller, and a 65% reduction in the worst-case control effort.