This study is concerned with the adaptive neural network (NN) observer design problem for continuous-time switched systems via quantized output signals. A novel NN presented in which laws are constructed using measurements. Then, persistent dwell time (PDT) switching considered to describe fast and slow a unified framework. Accurate estimations of state actuator efficiency factor can be obtained by proposed technique despite degradation. Finally, simulation example provided illustrate...

This paper is concerned with the fault reconstruction observer design problem unknown nonlinearities, external disturbances and faults. First, a neural-network-based estimation approach developed to generate estimations of actuator failures. In this design, neural network strategy utilized approximate totally nonlinear functions. Then, an iterative adaptive designed offer accurate sensor faults, where in previous iteration are applied current guarantee convergence errors. The can reconstruct...

This article addresses the simultaneous actuator and sensor fault estimation (FE) problem for a class of Markovian jump systems (MJSs) with nondifferentiable failures. In order to overcome difficulties brought by failures, we construct an extended vector composed states, faults, disturbances, where derivatives failures are not required in this augmented system. Then, two novel observer-based approaches developed descriptor system cope FE problem. The first one is reduced-order observer, can...

This article investigates the state and fault estimation problem for switched systems with external disturbances, unknown nonlinear dynamics, actuator sensor faults. In this article, a new adaptive fuzzy-based observer is designed, where prior knowledge of nonlinearities, faults not required. contrast to sliding mode (SMO), proposed can guarantee convergence error trajectory through online mechanism, avoiding surface switching in SMO design. The accurate asymptotic estimations also be...

This article investigates the fault reconstruction problem for a class of continuous-time switched systems subject to persistent dwell-time (PDT) switching. In order relax rank condition perturbations' coefficient matrices in fuzzy results, novel adaptive design approach is developed by decomposing perturbations into virtual faults and states. Then, PDT switching first considered observer overcome restricted frequency problem, which can guarantee stability estimation error with fast slow...

This article studies the fault estimation problem for a class of continuous-time nonlinear Markovian jump systems with unmeasured states, unknown bounded sensor faults, and nonlinearities simultaneously. In this article, new adaptive fuzzy observer design scheme is developed, where completely terms are approximated by logic systems. By means novel online mechanism, asymptotic stability error dynamic system guaranteed despite faults terms. Moreover, sliding surface switching in traditional...

This article investigates the fault estimation (FE) problem for a class of nonlinear systems via an adaptive fuzzy approach. Considering limited communication capacity networks, quantized measurement signals are used to construct laws instead real measurements in designed observer. By injecting quantizer parameter into observer inputs, quantization effects on convergence errors can be compensated. It is also shown that nondifferentiable actuator faults reconstructed by developed FE Finally,...

This paper concerns with the fault estimation problem for a class of unmanned marine vehicles sensor faults and non-differentiable actuator faults. Firstly, presented are modeled as continuous-time switched systems. In order to resolve faults, two kinds learning observer methods proposed allow augmented systems independent time-derivative The first one is kind timeline-based observer, which produces satisfactory results by using estimated information from previous moment measured output...

In this study, a thermal model of space telescope is established in Simulink. An intelligent autonomous control strategy based on actor-critic reinforcement learning (RL) for proportional...

With the remarkable capabilities, large language models (LLMs) have emerged as essential elements in numerous NLP applications, while parameter-efficient finetuning, especially LoRA, has gained popularity a lightweight approach for model customization. Meanwhile, various dropout methods, initially designed full finetuning with all parameters updated, alleviates overfitting associated excessive parameter redundancy. Hence, possible contradiction arises from negligible trainable of LoRA and...

The rapid evolution of large language models (LLMs) has transformed the competitive landscape in natural processing (NLP), particularly for English and other data-rich languages. However, underrepresented languages like Cantonese, spoken by over 85 million people, face significant development gaps, which is concerning given economic significance Guangdong-Hong Kong-Macau Greater Bay Area, substantial Cantonese-speaking populations places Singapore North America. Despite its wide use,...

In uncertain marine environment, auxiliary sensors of the unmanned vehicle (UMV) integrated navigation system may be abnormal at any time, reducing accuracy. To address this problem, article presents a novel data-based anomaly location and recovery (ALR) algorithm for strapdown inertial (SINS)/Doppler velocity log (DVL)/pressure sensor (PS) system. The ALR uses long short-term memory (LSTM) networks to establish relationship between filter parameters anomalies. Considering dependence...

Abstract Offset joints are widely used in robotics, and literature has demonstrated that axial offset can expand the workspace. However, hybrid joint, which incorporates offsets three orthogonal directions (x, y, z axes), provides a more flexible comprehensive range of motion compared to traditional joints. Therefore, understanding workspace with three-directional is essential. Firstly, through parameter model, interference studied, considering different directional obtaining analytical...

In order to solve the tracking accuracy problem of redundant manipulator, a PI control method with Henry gas solubility optimization parameter regulator (PI-HGSO) is proposed in this paper. This consists controller and regulator. The characteristic that position deviation manipulator equivalent specific function; namely, proportional-integral (PI) used adjust input. error can be better corrected by processing so inverse kinematics solution minimum realized. At same time, selection...

Suction hopper dredgers are the main type of ship used in dredging projects. Due to complex dynamic characteristics process and fact that power consumption affected by many factors such as soil, meteorological environment, hydrology, it is difficult obtain accurate prediction results using traditional methods. This paper proposes a data extraction method based on correlation analysis filter out most critical influencing features from large amount affects working efficiency dredgers. An...

This paper develops a fault tolerant control method for unmanned surface vehicle with unknown nonlinearity, actuator degradation and external disturbance simultaneously. The sliding mode controller is designed first to eliminate the effects of disturbances, where quantization errors between remote station can be compensated. Then, adaptive mechanism updated online applied deal nonlinearity degradation. Furthermore, proposed law shown guarantee strict stability reachability surface. Finally,...