Kemal Ucak

Adaptive PID controller based on online LSSVR with kernel tuning

In this paper, the effects of tuning the kernel bandwidth for an online LSSVM are investigated. LSSVM is used to obtain a model of the system, and based on this model information, an adaptive PID is designed to control the plant. The kernel parameter determines how the measured input is mapped to the feature space and a better plant model can be achieved by discarding redundant or irrelevant features, therefore introducing adaptability in kernel parameters improves modeling performance. The purpose of this paper is to find the optimal kernel bandwidth to improve the modeling performance of the LSSVM and consequently control performance obtained by adaptive PID which is designed based on the Jacobian information attained by the LSSVM. The proposed method has been evaluated by simulations carried out on a continuously stirred tank reactor (CSTR), and the results show that there is an improvement both in modeling and control performances.

An improved adaptive PID controller based on online LSSVR with multi RBF Kernel tuning

In this paper, the effects of using multi RBF kernel for an online LSSVR on modeling and control performance are investigated. The Jacobian information of the system is estimated via online LSSVR model. Kernel parameter determines how the measured input is mapped to the feature space and a better plant model can be achieved by discarding redundant features. Therefore, introducing flexibility in kernel function helps to determine the optimal kernel. In order to interfuse more flexibility to the kernel, linear combinations of RBF kernels have been utilized. The purpose of this paper is to improve the modeling performance of the LSSVR and also control performance obtained by adaptive PID by tuning bandwidths of the RBF kernels. The proposed method has been evaluated by simulations carried out on a continuously stirred tank reactor (CSTR), and the results show that there is an improvement both in modeling and control performances.

An adaptive support vector regressor controller for nonlinear systems

In this study, a novel online support vector regressor (SVR) controller based on system model estimated by a separate online SVR is proposed. The main idea is to obtain an SVR controller based on an estimated model of the system by optimizing the margin between reference input and system output. For this purpose, “closed-loop margin” which depends on tracking error is defined, then the parameters of the SVR controller are optimized so as to optimize the closed-loop margin and minimize the tracking error. In order to construct the closed-loop margin, the model of the system estimated by an online SVR is utilized. The parameters of the SVR controller are adjusted via the SVR model of system. The stability of the closed-loop system has also been analyzed. The performance of the proposed method has been evaluated by simulations carried out on a continuously stirred tank reactor (CSTR) and a bioreactor, and the results show that SVR model and SVR controller attain good modeling and control performances.

A trajectory tracking application of redundant planar robot arm via support vector machines

In this paper we present a kinematic based trajectory tracking application of redundant planar robot arm by using support vector machine method (SVM). The main advantages of using the proposed method are that, it does not suffer from singularity that is the main problem of redundancy in robot kinematics and better results for the kinematic model of redundant robot arm can be obtained by using less training data. Training data are obtained by using the forward differential kinematic model of the robot arm. We also implement the trajectory tracking application by using Artificial Neural Networks (ANN). Two methods are compared with respect to their generalization performances, and training performance. Simulation results are given.

Generalized self-tuning regulator based on online support vector regression

This paper introduces a novel generalized self-tuning regulator based on online support vector regression (OSVR) for nonlinear systems. The main idea is to approximate the parameters of an adaptive controller by optimizing the regression margin between reference input and system output. For this purpose, “closed-loop margin” which depends on tracking error is defined, and the parameters of the adaptive controller are optimized so as to minimize the tracking error which leads simultaneously to the optimization of the closed-loop margin. The overall architecture consists of an online SVR which computes a forward model of the system, an adaptive controller with tunable parameters and an adaptation mechanism realized by separate online SVRs to estimate each tunable controller parameter. The proposed architecture is implemented with adaptive proportional–integral–derivative (PID) and adaptive fuzzy PID in the controller block. The performance of the generalized self-tuning regulator mechanism has been examined via simulations performed on a bioreactor benchmark system, and the results show that the generalized adaptive controller and OSVR model attain good control and modeling performances.

A Novel Adaptive NARMA-L2 Controller Based on Online Support Vector Regression for Nonlinear Systems

In this study, a novel nonlinear autoregressive moving average (NARMA)-L2 controller based on online support vector regression (SVR) is proposed. The main idea is to obtain a SVR based NARMA-L2 model of a nonlinear single input single output system (SISO) by decomposing a single SVR which estimates the nonlinear autoregressive with exogenous inputs (NARX) model of the system. Consequently, using the obtained SVR-NARMA-L2 submodels, a NARMA-L2 controller is designed. The performance of the proposed SVR based NARMA-L2 controller has been evaluated by simulations carried out on a bioreactor system, and the results show that the SVR based NARMA-L2 model and controller attain good modelling and control performances. Robustness of the controller in the case of system parameter uncertainty and measurement noise have also been examined.

Controlling 3-DOF helicopter via fuzzy PID controller

In this paper, the real time tracking performance of the fuzzy PID controller has been evaluated on nonlinear 3-DOF helicopter system. Quanser 3-DOF helicopter system has been utilized as experimental setup. Three separate fuzzy PID controller have been employed so as to control three state of the system. The dynamics of the system has been altered via counterweight in order to examine the performances of the both controller for parametric uncertainty in system parameter. The results reveal that the fuzzy PID controller leads to a good tracking performance and has better parametric uncertainty rejection capability than LQR.

Fault diagnosis in a nonlinear three-tank system via ANFIS

In this paper, two intelligent methods namely Artificial Neural Network (ANN) and Adaptive Neuro Fuzzy Inference Systems (ANFIS), are implemented to diagnose the leakage faults in a nonlinear three tank system. Two separate structures are utilized for fault diagnosis. One is to identify the dynamics of the plant and the other is to construct the residual logic mechanism. The performance of the proposed methods are evaluated by simulations carried out on a three tank system (TTS). The leakages in tanks are considered as faults in the tank system.

Intelligent systems based solutions for the kinematics problem of the industrial robot arms

In this paper, three intelligent system methods namely Artificial Neural Network (ANN), Support Vector Machine (SVM) and Adaptive Neuro Fuzzy Inference Systems (ANFIS), are implemented to solve the inverse kinematics problem of the industrial robot arms. The main advantages of the intelligent system based solutions in the robot kinematics are that they can be easily implemented in analysis of complex mechanisms and their solutions do not suffer by the singularity that is one of the fundamental problems of inverse kinematics. The screw theory and quaternion algebra based kinematic model is used to improve the model efficiency by decreasing the computational complexity and load. The kinematics problem of the Staubli TX-60L industrial robot arm is analyzed by using the proposed intelligent system based solutions and simulation results are given.

Train Speed Control in Moving-Block Railway Systems: An Online Adaptive PD Controller Design

Abstract Since the overall capacity of railway lines are not used effectively in fixed-block railway systems, moving-block railway systems are introduced to increase the transport capacity and to reduce the headway times. CBTC (Communication Based Train Control) systems and ERTMS (European Rail Traffic Management System) application level 3 are regarded as examples of moving-block systems. In this study, the concept of moving-block system is explained and an adaptive PD control technique based on Online Least Square Support Vector Regression is used for train speed control.