The dual neural extended Kalman filter (DNEKF) is proposed in this article to compensate for model inaccuracies and violations of noise assumption the EKF-based multirate sensor fusion. DNEKF employs two networks improve process state output predictions through simultaneous parameter estimations using vector augmentation. Taking advantage frequent but less accurate measurements infrequent more provides a update strategy training approach. Two numerical examples one industrial application...

Sensor fusion plays a critical role in improving estimation accuracy of process quality variables. In this article, the dual, neural, extended Kalman filter (DNEKF) and state model compensation (SNEKF) are synthesized to compensate for modeling errors (EKF)-based multirate sensor fusion. Specifically, is performed presence irregularly sampled, slow-rate measurements with time-varying time delays. The proposed algorithm estimates neural network parameters simultaneously through vector...

Genetic algorithms (GAs) are intelligent computational tools which their simplicity, accuracy and adaptable topology cause them to be used in globally minimum or maximum finding problems. Developing the GAs increase speed global of a cost function has been big challenge until now many variants GA evolved accomplish this goal. This paper presents two new sequential mutation method circular gene GA. These methods attain better final answer accompanied by lesser use evaluations comparison with...

In this study, a generalised minimum variance control (GMVC) method using the projection-based recurrent neural network (PRNN) is proposed to minimise error in output of non-linear plant. One main drawbacks conventional GMVC approaches lack systematic procedure deal with input constraints. PRNN employed for incorporating constraints index. This based on optimality conditions constrained problem and designed projection theorem. To formulate approach, by considering an ARMAX model system...

Genetic algorithm and particle swarm optimization are two methods which can be used to find the global extremum of cost functions. The solely performance each method their specific characteristics in finding have been giving idea hybridization these many researchers. In this paper a new hybrid named Serial Algorithm Particle Swarm Optimization (SGAPSO) is introduced configuration discussed details. A set benchmark functions consisted high dimensional, multimodal low dimensional compare...

The Minimum Variance Lower Bound (MVLB) represents the best achievable controller capability in a variance sense. Estimation of MVLB for nonlinear systems confronts some difficulties. If one simply ignores these nonlinearities, there is danger over-estimating performance control loop rejecting uncertainties. Assuming that almost all models have uncertainties, this paper, has been estimated considering three types uncertainties: structural, parametric, and algorithmic. To achieve accurate...

To assess the performance of a control loop based on minimum variance (MV) benchmark, we need to calculate MV lower bound (MVLB). Even though there is plethora literature available for calculating MVLB linear systems, these methods are not suitable non‐linear systems. Furthermore, almost all real‐world applications have been encountered with input constraints. These constraints limit controllers' abilities in decreasing output variability. Therefore, existing computation methods, which do...

This paper presents a control strategy for achieving robust minimum variance controller (MVC) by modelling uncertainty using type-2 fuzzy set and satisfying H ∞ norm specifications. In this paper, an MVC is designed considering three types of structural, parametric algorithmic uncertainties. Thus, interval utilized. utilizes one method to construct models the uncertain data model error approach. addition, sufficient condition developed that guarantees stability robustness controller. The...

Abstract This study presents a new fault detection scheme based on the probability density function (PDF) of system output. Unlike classical and diagnosis methods, in proposed method, distribution output is estimated online. To achieve this goal, an algorithm introduced to estimate PDF online using fuzzy logic. Furthermore, convergence investigated. Then, residual constructed that can show existence system. The main advantages method are robustness against measurement noise, even though it...

Summary This paper is concerned with distributed data‐driven observer design problem. The existing observers rely on a common assumption that all the information about system, and calculations based upon this are centralized. Therefore resulting algorithms cannot be applied to systems in which each local receives only part of output signal. On other hand, traditional model‐based state estimation methods generally assume processes decomposed according known process models, while approaches...

This paper deals with the control performance assessment of distributed networked systems in presence communication delays. The best achievable steady-state framework linear quadratic Gaussian (LQG) is presented where delays serve as most fundamental limitation. In proposed LQG design, controllers and observers are designed simultaneously without using separation principle. And non-applicability principle investigated demonstrated order to handle time-varying delays, we propose use lower...