ISA transactions最新文献

This study proposes a novel pressure control strategy for a centrifugal fan driven by an induction motor, using a two-degree-of-freedom (2-DOF) proportional-integral (PI) controller optimized via the fungal growth optimizer (FGO). Built upon a validated nonlinear model, the FGO algorithm tunes the controller parameters to minimize a cost function reflecting overshoot and tracking error. Comparative evaluations are performed against genetic algorithm (GA), L-SHADE, and the Newton-Raphson-based optimizer (NRBO), as well as classical methods such as Cohen-Coon and Rovira techniques. The FGO-based controller demonstrates superior dynamic response, robustness, and noise immunity. Statistical validation using the Wilcoxon signed-rank test confirms the performance improvements. These results highlight the effectiveness of FGO in real-world pressure regulation and establish a benchmark for modern industrial fan control systems.

To enhance the robustness and control efficiency of underwater robot systems, this paper presents a controller that integrates model predictive control (MPC) with a prescribed time control, and introduces a novel time controller capable of more intuitively and effectively determining the convergence time of system performance. A radial basis function neural network (RBFNN) coupled with an evaluation neural network has been engineered to bolster the system's capacity to address ocean current disturbances. The introduction of a new error function ensures the effective realization of the neural network's stability. By incorporating a MPC outer loop controller with system state parameters, the system state can be transitioned smoothly and kept within bounds, aligning with the actual performance characteristics of underwater vehicles. The system's stability is verified through theoretical analysis and further substantiated through simulation experiments. Additionally, the performance enhancement of the robot system achieved by the proposed algorithm is emphasised through a comparative analysis with the traditional PID control algorithm. Ultimately, in light of the complex, time-varying external disturbance environment delineated in this paper, a sea trial experiment is conducted to ascertain the effectiveness of the proposed algorithm.

In this article, a fault-tolerant adaptive tracking control issue is investigated for nonaffine fractional-order nonlinear systems (FONSs) with prescribed-time interval output constraints (PTIOCs), nonlinear input and actuator fault. The PTIOCs mean that the output constraints begin sometime after system operation and have a prescribed-time duration. It should be emphasized that both the start moment and the end moment of the constraints can be determined by the user. Using the improved dependent-constrained-error function and prescribed-time scaling function, as well as with the help of a barrier function, the PTIOCs issue is transformed into verifying the boundedness of the barrier function. Moreover, an adaptive fault-tolerant control algorithm is developed to mitigate the effects of actuator faults in FONSs while accommodating multiple constraints and nonlinear inputs without requiring structural changes to the controller. The validity of the proposed scheme is then substantiated by simulations.

In this paper, a new control approach is proposed for the synchronization and stabilization of a class of chaotic systems with unknown dynamics and input nonlinearities. Type-3 fuzzy logic systems (T3-FLSs) are developed to adaptively model the dynamics of both master and slave systems in real time. The input is affected by sector-bounded hysteresis and quantization, and these challenges are explicitly addressed in the control design. Unlike conventional methods, the proposed strategy does not require prior knowledge of the system equations or the derivatives of system signals. The adaptation laws for the T3-FLS parameters and estimation errors are rigorously derived using stability and robustness analysis, ensuring smooth control signals without chattering. Extensive simulations and real-time examinations demonstrate that the method achieves accurate synchronization even under severe uncertainties, high levels of random noise, and non-identical chaotic systems. Comparative results confirm the superiority of the proposed approach over existing fuzzy control methods.

In this paper, a zonotopic space interpolation-based smooth variable structure filter (ZSI-SVSF) is proposed for state estimation under unknown but bounded noise. The approach combines zonotopic contraction to determine a state confidence envelope with cubic spline interpolation to construct a smooth boundary layer. A curvature-driven mechanism then directs the smooth variable structure filter for iterative updates, adaptively computing the estimation gain and yielding a smooth trajectory. Unlike conventional filters, ZSI-SVSF avoids covariance computations, improving efficiency, reducing conservativeness, and mitigating chattering. Its applicability extends from linear systems to nonlinear measurements via Jacobian linearization. Validation through linear simulation with a hybrid motion model and nonlinear UAV tracking under a coordinated turn model confirms improved accuracy, adaptability, and low computational cost.

Hysteresis nonlinearities, time delays and output constraints are widely present in the practical physical systems such as motion platforms driven by smart material actuators, bionic robots and ultra-high precision machining systems, however, these factors will degrade the control performance and may even induce oscillations in the control system. To overcome the aforementioned problems, in this paper, an adaptive neural pseudo-inverse control scheme is proposed for a class of time-delay nonlinear hysteresis systems considering output constraints with the following contributions: 1) a novel butterfly-like Krasnoselskii-Pokrovskii (BKP) hysteresis model with double loops is constructed to describe the double-loop hysteresis by performing the weighted superposition of the new proposed butterfly-like KP kernel; 2) a novel adaptive pseudo-inverse control algorithm is developed to avoid the difficulty of constructing the direct double-loop inverse model; 3) a new motion control platform actuated by the flexible dielectric elastomer actuators is established to verify the effectiveness of the proposed control scheme and demonstrate its feasibility for drive control systems in soft bionic robots.

Recognizing nonstationarity is pivotal for trustworthy industrial process monitoring. Existing methods address this issue from a unimodal perspective, which struggles to capture intrinsic heterogeneity. To resolve this, we introduce a novel unsupervised multimodal nonstationary monitoring framework (UMNMF), integrating a bimodal paradigm with contrastive and adversarial schemes. Initially, the knowledge labeling unit (KLU) is established to generate pseudo-labels augmented with prior knowledge for semantic guidance. Subsequently, the dynamic alignment and encoding unit (DAEU) exploits contrastive language-image pre-training (CLIP) and the Vision Transformer (ViT) for modality-aware alignment through a pseudo-supervised contrastive mechanism. Furthermore, the association alignment and distillation unit (AADU) is devised to achieve decoupling through self-adversarial distribution regularization within a variational graph autoencoder (VGAE). The superior performance is substantiated by extensive experiments on three industrial processes, where the UMNMF attains an average fault detection rate exceeding 94 % and maintains a false alarm rate below 2.5 %. Additional ablation studies further confirm the contribution of each module to overall performance improvement.

In practice, actuators inherently have natural limitations that are often overlooked by theoretical control strategies, such as bounded output and discrepancies between calculated and actual responses. Consequently, reliable control techniques must be employed to ensure safe and dependable operation. We propose an adaptive control law for global regulation in the case of partial loss of effectiveness due to input faults. The proposed control strategy maintains the system at the desired configuration while accounting for the physical limitations of the actuators. Additionally, it is suitable for dealing with unknown parameters in the vector of gravitational forces, making it robust to these uncertainties. The approach treats the inputs as time-varying signals, allowing for an infinite number of faults. The methodology has the advantage of operating in continuous time. Thus, it is theoretically capable of counteracting any fault event of arbitrary magnitude, as long as the actuators are able to continue overcoming gravitational effects while maintaining inertial and gravitational counteracting responses. The closed-loop system is analyzed using Lyapunov's stability theory for non-autonomous systems, concluding global asymptotic convergence of all signals. The effectiveness of the control strategy is illustrated through simulations that exhibit resilience to degradation in actuator performance.

Regarding the accurate and fast-tracking control problem of a multi-joint manipulator system under the multiple uncertainties of actuator faults, parameter perturbations, and external disturbances, this paper proposes fault-tolerant control scheme based on a predefined time Extended state observer. First, a novel extended state observer has been designed to estimate the time-varying composite disturbance within a predefined time. Second, a new predefined time integral sliding mode surface and Finite time prescribed performance constraint function are constructed, and a continuous function is introduced to reduce chattering. The Lyapunov stability theory is used to analyze the predefined time stability characteristics of the closed-loop system. Finally, numerical simulations verify the effectiveness of the proposed control method, and the system has strong robustness and high steady-state accuracy, and can withstand actuator failures and other uncertain influencing factors.

Existing safe backstepping control methods often rely on precise system models, which limits their robustness against external disturbances frequently encountered in real-world applications. To address the problem of safe control for strict-feedback nonlinear systems subject to unknown disturbances, this paper proposes a novel constructive control framework that integrates a disturbance observer, filtered backstepping, and control barrier functions (CBFs). A disturbance observer is first developed to estimate and compensate for unknown disturbances, providing a less conservative alternative to conventional robust methods that rely on "worst-case" assumptions. The estimated disturbance is then embedded into a recursive design framework that combines filtered backstepping with CBFs to construct smooth virtual and actual controllers. This unified approach mitigates the "explosion of complexity" typically associated with safe backstepping while avoiding the nonsmoothness inherent in optimization-based CBF controllers. Finally, the effectiveness of the proposed method is validated through numerical simulations and real-time experiments on a Franka Emika Panda robotic arm.