Sharif Digital Repository

Deep neural networks have presented superlative performance in many machine learning based perception and recognition tasks, where they have even outperformed human precision in some applications. However, it has been found that human perception system is much more robust to adversarial perturbation, as compared to these artificial networks. It has been shown that a deep architecture with a lower Lipschitz constant can generalize better and tolerate higher level of adversarial perturbation. Smooth regularization has been proposed to control the Lipschitz constant of a deep architecture and in this work, we show how a deep convolutional neural network (CNN), based on non-smooth regularization... 

This article focuses on the robust D-stabilization analysis of fractional-order control systems where each of the system and the controller may be of fractional order. The coefficients of the system are considered as complex linear functions of interval uncertain parameters, so this article deals with fractional-order polytopic systems. First, a necessary and sufficient condition is introduced for the robust D-stabilization of the closed-loop control system based on the zero exclusion condition and the value set concept. Then, the geometric pattern of the value set of the characteristic polynomial is obtained analytically using the exposed vertices. Second, a function is presented to check... 

Fractional order PID controllers are suitable for almost all types of dynamic models. In this paper, a new adaptive fractional order PID controller using neural networks is introduced.The overall performance using the proposed adaptive fractional order PID controller is demonstrated through some examples. It is shown that the new controller scheme can give excellent performance and more robustness in comparison with the conventional controllers like GPC  

The problem of time optimal magnetic attitude control is treated and an open loop solution is first obtained using a variational approach. In order to close the control loop, a neural network with time varying weights is proposed as a feedback optimal controller applicable to the time varying nonlinear system. The good robustness and low real-time computational burden of the proposed neuro-controller makes the controller more useful compared to the other control methods. © 2008 IEEE  

An economic-statistical design with multiple assignable causes following exponential distribution is presented in this paper for linear profiles. For this purpose, a tri-objective optimization model is proposed to minimize the cost with desired statistical performances. Average Run Length (ARL) as the primary statistical measure is employed for the appraisal of the designed linear profiles. The first objective to be minimized is a cost function that models the implementation cost in different states. The second objective is to maximize ARL or the in-control average run-length of the monitoring scheme. The third objective to be minimized is ARL 1 or the out-of-control average run-length of... 

Reducing the renewable energy costs is necessary for the competition with the fossil energies and control strategies have great impact on the efficiency of wind machines. In the wind turbine industry, a practical approach is to maximize the energy capture of a wind machine by optimizing the power coefficient in the under-rated situations. In this paper, with the main objective of maximizing the energy capture in the second region, four different control strategies are compared in the presence of uncertainties. The proposed control methods are compared based on their power capture and robustness against probable uncertainties in the structural and environmental parameters. A two-mass... 

This letter presents a unified framework for the design of prescribed-time controllers under time-varying input and state constraints for normal-form unknown nonlinear systems with uncertain input gain. The proposed approach is based on a time-domain mapping method by which any infinite-time system can be corresponded to a prescribed-time system and vice versa. It is shown that the design of a constrained nonasymptotic prescribed-time controller can be reduced to the asymptotic control design for an associated constrained infinite-time system. Faà di Bruno’s formula and Bell polynomials are used for a constructive representation of the associated infinite-time system. The presented results... 

This article focuses on the robust D-stabilization analysis of fractional-order control systems where each of the system and the controller may be of fractional order. The coefficients of the system are considered as complex linear functions of interval uncertain parameters, so this article deals with fractional-order polytopic systems. First, a necessary and sufficient condition is introduced for the robust D-stabilization of the closed-loop control system based on the zero exclusion condition and the value set concept. Then, the geometric pattern of the value set of the characteristic polynomial is obtained analytically using the exposed vertices. Second, a function is presented to check... 

Reducing the renewable energy costs is necessary for the competition with the fossil energies and control strategies have great impact on the efficiency of wind machines. In the wind turbine industry, a practical approach is to maximize the energy capture of a wind machine by optimizing the power coefficient in the under-rated situations. In this paper, with the main objective of maximizing the energy capture in the second region, four different control strategies are compared in the presence of uncertainties. The proposed control methods are compared based on their power capture and robustness against probable uncertainties in the structural and environmental parameters. A two-mass... 

A multi-objective economic-statistical design is aimed in this article for simple linear profiles. In this design, the interval between two successive sampling intervals, the sample size and the number of adjustment points alongside, the parameters of the monitoring scheme are determined such that not only the implementation cost is minimized, but also the profile exhibits desired statistical performances. To this aim, three objective functions are considered in the multi-objective optimization model of the problem. The Lorenzen–Vance cost function is used to model the implementation cost as the first objective function to be minimized. The second objective function maximizes the in-control... 

This letter presents a unified framework for the design of prescribed-time controllers under time-varying input and state constraints for normal-form unknown nonlinear systems with uncertain input gain. The proposed approach is based on a time-domain mapping method by which any infinite-time system can be corresponded to a prescribed-time system and vice versa. It is shown that the design of a constrained nonasymptotic prescribed-time controller can be reduced to the asymptotic control design for an associated constrained infinite-time system. Faà di Bruno's formula and Bell polynomials are used for a constructive representation of the associated infinite-time system. The presented results... 

In this research, the main parameters of the multivariate cumulative sum (CUSUM) control chart (the reference value k, the control limit H, the sample size n and the sampling interval h) are determined by minimising the Lorenzen-Vance cost function [Lorenzen, T.J., and Vance, L.C. (1986), The Economic Design of Control Charts: A Unified Approach, Technometrics, 28, 3-10], in which the external costs of employing the chart are added. In addition, the model is statistically constrained to achieve desired in-control and out-of-control average run lengths. The Taguchi loss approach is used to model the problem and a genetic algorithm, for which its main parameters are tuned using the response... 

In this research, an iterative approach is employed to analyze and classify the states of uni-variate quality control systems. To do this, a measure (called the belief that process is in-control) is first defined and then an equation is developed to update the belief recursively by taking new observations on the quality characteristic under consideration. Finally, the upper and the lower control limits on the belief are derived such that when the updated belief falls outside the control limits an out-of-control alarm is received. In order to understand the proposed methodology and to evaluate its performance, some numerical examples are provided by means of simulation. In these examples, the... 

Application of fractional order PID (FOPID) controller to an automatic voltage regulator (AVR) is presented and studied in this paper. An FOPID is a PID whose derivative and integral orders are fractional numbers rather than integers. Design stage of such a controller consists of determining five parameters. This paper employs particle swarm optimization (PSO) algorithm to carry out the aforementioned design procedure. PSO is an advanced search procedure that has proved to have very high efficiency. A novel cost function is defined to facilitate the control strategy over both the time-domain and the frequency-domain specifications. Comparisons are made with a PID controller and it is shown... 

Predictive functional control (PFC) as a new active queue management (AQM) method in dynamic TCP networks supporting explicit congestion notification (ECN) is proposed. The ability of the controller in handling system delay along with its simplicity and low computational load makes PFC a privileged AQM method in the high speed networks. Besides, considering the disturbance term (which represents model/process mismatches, external disturbances, and existing noise) in the control formulation adds some level of robustness into the PFC-AQM controller. This is an important and desired property in the control of dynamically-varying computer networks. In this paper, the controller is designed based... 

In most real-world manufacturing systems, the production of goods comprises several autocorrelated stages and the quality characteristics of the goods at each stage are correlated random variables. This paper addresses the problem of monitoring a multivariate-multistage manufacturing process and diagnoses the possible causes of out-of-control signals. To achieve this purpose using multivariate time series models, first a model for the autocorrelated data coming from multivariate-multistage processes is developed. Then, a single neural network is designed, trained and employed to control and classify mean shifts in quality characteristics of all stages. In-control and out-of-control average... 

Haptic interfaces, a kinesthetic link between a virtual environment and a human operator play a pivotal role in the reproduction of realistic haptic force feedback of the virtual reality-based simulators. Since most of the practical control theories are model-based, the identification of the robot’s dynamics, for precise modeling of the system dynamics, is a process of high significance and usage. This research addresses dynamic characterization, performance issues, and structural stability, associated with a parallel haptic device interaction with an admittance type virtual environment. In this regard, considering the Lion identification scheme, we characterized the dynamics of a robot... 

In this paper we follow two approaches in optimal nonlinear control of a snake-like robot. After deriving the dynamic equations of motion using Gibbs-Appell method, reducing these equations, and some assumptions, feedback-linearization method was used to linearize the nonlinear system. The obtained controller is used in simulations to control robot to track a desired line, with minimum required torques. Two goals are desired. First the robot's head is expected to track a distinct line with a given speed. And next, tracking the serpenoid curve is desired. The simulation results prove the controller efficiency. The robustness of the designed controller is shown by comparing the torques with... 

Integrating humans and robotics technology into one system offers multiple opportunities for creating assistive technologies that can be employed in biomechanical, industrial, and aerospace applications. The present study deals with employing the adaptive neuro-fuzzy inference system (ANFIS) approach in rule base derivation for powered exoskeleton intelligent control to assist paraplegic patient mobility. By employing the hybrid learning algorithm, optimal distributed membership functions to describe the mapping relation in the input and output parameters of the gait cycle were derived. As the proposed control strategy was unaffected by changing human dynamics, the reliability and robustness... 

So the design and control of an accurate robot for this purpose is very critical for saving the patients. Modification of the model and designing two optimized nonlinear robust controllers for the first time for the parallel manipulator medical robot and cardiopulmonary resuscitation. The main objective of the current study in order to decrease the overshoot and increase the accuracy of the position and convergence speed and robustness to destructive factors affecting the precision of the robot. In this paper firstly, the kinematics and dynamics analysis of a translational parallel manipulator robot is presented and a non-linear model in the presence of uncertainties, disturbances, and...