Mohammad Mansouri

Load frequency control in interconnected power system using modified dynamic neural networks

This paper present power system load frequency control by modified dynamic neural networks controller. The controller has dynamic neurons in hidden layer and conventional neurons in other layers. For considering the sensitivity of power system model, the neural network emulator used to identify the model simultaneously with control process. To have validation of proposed structure of neural network controller the results of simulation demonstrated that the proposed controller offers better performance than conventional neural network controller.

Average-inertia weighted cat swarm optimization

For improving the convergence of Cat Swarm Optimization (CSO), we propose a new algorithm of CSO namely, Average-Inertia Weighted CSO (AICSO). For achieving this, we added a new parameter to the position update equation as an inertia weight and used a new form of the velocity update equation in the tracing mode of algorithm. Experimental results using Griewank, Rastrigin and Ackley functions demonstrate that the proposed algorithm has much better convergence than pure CSO.

Control of Flexible Joint Manipulator via Reduced Rule-Based Fuzzy Control with Experimental Validation

A novel structure of fuzzy logic controller is presented for trajectory tracking and vibration control of a flexible joint manipulator. The rule base of fuzzy controller is divided into two sections. Each section includes two variables. The variables of first section are the error of tip angular position and the error of deflection angle, while the variables of second section are derivatives of mentioned errors. Using these structures, it would be possible to reduce the number of rules. Advantages of proposed fuzzy logic are low computational complexity, high interpretability of rules, and convenience in fuzzy controller. Implementing of the fuzzy logic controller on Quanser flexible joint reveals efficiency of proposed controller. To show the efficiency of this method, the results are compared with LQR method. In this paper, experimental validation of proposed method is presented.

Training ANFIS system with DE algorithm

In this study, a new type of training the adaptive network-based fuzzy inference system (ANFIS) is presented by applying different types of the Differential Evolution branches. The TSK-type consequent part is a linear model of exogenous inputs. The consequent part parameters are learned by a gradient descent algorithm. The antecedent fuzzy sets are learned by basic differential evolution (DE/rand/1/bin) and then with some modifications in it. This method is applied to identification of the nonlinear dynamic system, prediction of the chaotic signal under both noise-free and noisy conditions and simulation of the two-dimensional function. Instead of DE/rand/1/bin, this paper suggests the complex type (DE/current-to-best/1+1/bin & DE/rand/1/bin) on predicting of Mackey-glass time series and identification of a nonlinear dynamic system revealing the efficiency of proposed structure. Finally, the method is compared with pure ANFIS to show the efficiency of this method.

Integer GA for mobile robot path planning with using another GA as repairing function

In this study integer genetic algorithm is applied for path planning of mobile robot in the grid form environment. The novel representation is proposed for definition of chromosome which reduced the computational complexity of genetic algorithm which was used before for path planning. Comparison with other encoding of chromosome is done to show the capability of proposed algorithm. Another genetic algorithm is used to repair some paths which collide with obstacles. Mamadani fuzzy rule is used to describe difficulty of passing from cells which are sandy or have slope.

Face Recognition Using Convolutional Neural Network and Simple Logistic Classifier

In this paper, a hybrid system is presented in which a convolutional neural network (CNN) and a Logistic regression classifier (LRC) are combined. A CNN is trained to detect and recognize face images, and a LRC is used to classify the features learned by the convolutional network. Applying feature extraction using CNN to normalized data causes the system to cope with faces subject to pose and lighting variations. LRC which is a discriminative classifier is used to classify the extracted features of face images. Discriminant analysis is more efficient when the normality assumptions are satisfied. The comprehensive experiments completed on Yale face database shows improved classification rates in smaller amount of time.

HIERARCHICAL STRUCTURE BASED CONVOLUTIONAL NEURAL NETWORK FOR FACE RECOGNITION

In this paper, a hierarchical structure based convolutional neural network is proposed to provide the ability for robust information processing. The weight sharing ability of convolutional neural networks (CNNs) is considered as a level of hierarchy in these networks. Weight sharing reduces the number of free parameters and improves the generalization ability. In the proposed structure, a small CNN which is used for feature extractor is shared between the whole input image pixels. A scalable architecture for implementing extensive CNNs is resulted using a smaller and modularized trainable network to solve a large and complicated task. The proposed structure causes less training time, fewer numbers of parameters and higher test data accuracy. The recognition accuracy for recognizing unseen data shows improvement in generalization. Also presented are application examples for face recognition. The comprehensive experiments completed on ORL, Yale and JAFFE face databases show improved classification rates and reduced training time and network parameters.

Adaptive variable structure hierarchical fuzzy control for a class of high-order nonlinear dynamic systems

In this paper, a novel adaptive hierarchical fuzzy control system based on the variable structure control is developed for a class of SISO canonical nonlinear systems in the presence of bounded disturbances. It is assumed that nonlinear functions of the systems be completely unknown. Switching surfaces are incorporated into the hierarchical fuzzy control scheme to ensure the system stability. A fuzzy soft switching system decides the operation area of the hierarchical fuzzy control and variable structure control systems. All the nonlinearly appeared parameters of conclusion parts of fuzzy blocks located in different layers of the hierarchical fuzzy control system are adjusted through adaptation laws deduced from the defined Lyapunov function. The proposed hierarchical fuzzy control system reduces the number of rules and consequently the number of tunable parameters with respect to the ordinary fuzzy control system. Global boundedness of the overall adaptive system and the desired precision are achieved using the proposed adaptive control system. In this study, an adaptive hierarchical fuzzy system is used for two objectives; it can be as a function approximator or a control system based on an intelligent-classic approach. Three theorems are proven to investigate the stability of the nonlinear dynamic systems. The important point about the proposed theorems is that they can be applied not only to hierarchical fuzzy controllers with different structures of hierarchical fuzzy controller, but also to ordinary fuzzy controllers. Therefore, the proposed algorithm is more general. To show the effectiveness of the proposed method four systems (two mechanical, one mathematical and one chaotic) are considered in simulations. Simulation results demonstrate the validity, efficiency and feasibility of the proposed approach to control of nonlinear dynamic systems.

Identification of Hammerstein-Wiener ARMAX systems using Extended Kalman Filter

In this study, Extended Kalman Filter (EKF) algorithm is developed to estimate the parameters of Hammerstein-Wiener (H-W) ARMAX models. The basic idea is to estimate the original parameters of the identification model, which are appeared in the form of product terms, directly. While, other algorithms like Extended Forgetting Factor Stochastic Gradient (EFG), Extended Stochastic Gradient (ESG), Forgetting Factor Recursive Least Square (FFRLS) and Kalman Filter (KF), estimate parameters in the product form and they need another algorithms such averaging method (AVE method), singular value decomposition method (SVD method) to separate the parameters. So, the computational complexity of the proposed approach decreases. To show the efficiency of this method the results are compared with EFG and ESG method.

A review on applicability of expert system in designing and control of autonomous cars

This paper presents a survey on applicability of expert system in designing and control of autonomous vehicles. Each of reviewed papers categorized on five categories as expert system, fuzzy expert system, neuro-fuzzy expert system, genetic plus neuro-fuzzy expert system and transition systems. Some earlier works used only a rule based expert system for driving the car but due to the lack of a precise model because of the uncertainty in modelling and have a simple, rapid and easily maintained design some others used fuzzy rule based system. The rules extracted from an experts knowledge, because of some fuzzy bottlenecks some articles used combination of neural network or genetic algorithms with fuzzy to produce an efficient knowledge base and finally there is an article that used a transition system that is extracted from an expert knowledge to produce the rule base of the expert system. The goal of the expert system is to issue commands to the controllers to do the appropriate task. The review reveals that due to some advantages of fuzzy, neural and genetic algorithms, most of articles used from this techniques beside expert system.