Levent Acar

Design of knowledge-rich hierarchical controllers for large functional systems

A hierarchical structure that utilizes all the functionalities of a large-scale system and unifies the dynamics of the system with its functional behavior is introduced. The proposed hierarchy is formed by decomposing the physical structure of a system and by associating knowledge-rich controllers with the structure. The inclusion of structural information in the hierarchy has several advantages. First, it presents multifunctional descriptions of portions of the system. Then, it provides a modular decomposition such that complete reconstruction of the hierarchy is not required if some parts of the system change. Most importantly, it enables local failure handling and replanning. To demonstrate the physical decomposition, task assignment, and control process, a system with two robot arms and a camera was considered as an example. >

Some examples for the decentralized receding horizon control

In the control of industrial processes, the use of finite-horizon quadratic cost criteria is becoming more popular. One type of control, which is called the predictive or the receding horizon control, utilizes finite-horizon optimization methods to obtain an infinite-horizon control. In this type of control, the horizon is constantly rolled back, and the control is constantly updated to reflect the possible changes in the system. This approach is most suited for applications in which future values are unpredictable, unreliable, or time varying. Examples of a receding horizon control are considered for decentralized systems. The performance of the decentralized receding horizon control for time-invariant and time-varying systems with different updating intervals is examined.<<ETX>>

An analysis of type-1 and type-2 fuzzy logic systems

This document introduces the concept of type-2 fuzzy sets and compares the type-2 fuzzy set preliminaries with the ordinary type-1 fuzzy set. Two theorems are proved: the first one helps to generate the general form of the type-2 operations, the second provides an approach to transfer the type-1 fuzzy uncertainties to the type-2 system.

Real-time nonlinear optimal control using neural networks

In this paper, a neural network based controller which optimizes a finite horizon quadratic cost function is developed for a class of nonlinear systems. The controller converges to its optimal value real-time eliminating the need for a priori knowledge of the nonlinearity and the initial conditions. The method makes use of the optimality conditions obtained from the Hamiltonian directly. These conditions are realized by a series of neural networks which converge to the optimal control iteratively in real-time. A nonlinear system to demonstrate its applicability is also included.

A neural network based approach for the identification and optimal control of a cantilever plate

Outlines a neural network based identification and optimal control approach for a specific nonlinear system that consists of a cantilever plate. The neural networks employed are multi-layer perceptrons with backpropagation learning method. The identifier is implemented in time domain to represent system nonlinearities. The backpropagation method is chosen so that the Jacobian of the system dynamics can be acquired directly and utilized later in obtaining the optimal control. The controller is designed to minimize a finite horizon quadratic cost function by solving the Hamiltonian equations. In order to compensate for the error accumulation between the model and the real system, the receding horizon control method is implemented.

Design of hierarchically distributed expert controllers for large-scale systems

An architecture with expert distributed controllers to control large functional systems hierarchically is considered. A structural hierarchy was formed based on the physical structure of the system and functions associated with the structure. The hierarchy and the functional associations are described completely by the structural and functional coordinability axioms. These axioms display a very specific type of hierarchical structure. One of the properties of the hierarchy is that the whole system is represented both at the top and at the tip levels. It has the ability to represent the whole system at any desired detail. Moreover, representational redundancies are deliberately allowed, but the top and tip levels are forced to be represented without redundancies. The authors define the functional behavior of the system and the expert controller via six primitives. These primitives express the functions, capabilities, restrictions, and information to control the behavior of the system. Finally, the control process is described as a uniform flow from top to tip.<<ETX>>

A testbed system for nonlinear or intelligent control

There has been a large amount of publications concerning nonlinear and intelligent control in recent years. It has become an important field of research. As each new algorithm becomes available, the need for realistic and accurate test bed systems is apparent. Some previous simulation models have not been real world systems, or a simplification of a much larger system. It is the goal of this document to present a much more complete description of one real world model so that others may use this model to test their control algorithms. The system analyzed in this document is a trailer truck system. The model, based on the physical system, extends the models of previous publications. A full dynamical representation is given. Tests show the models accuracy against the actual system.

Neural network based control for a backward maneuvering trailer truck

The control of non-linear systems using neural networks has gained increasing interest in recent years. The non-linear capabilities of neural networks can be utilized for a large number of non-linear systems that are not controllable by linear techniques. However, controllers based on neural networks are quite difficult to design due to the lack of good training data. The physical plant developed and designed in this work is a scale model of a trailer truck. A neural network based controller is to be developed to drive the truck backwards from any initial condition to a loading dock. One approach to solve this control problem is to design a valid path, and then track the path. Since the valid path depends on initial conditions, a neural network is trained to generate selective points on the path for any initial condition. The path data used to train the neural network are obtained from software based on a previously published work on similar systems. Based on generated path data, a control is designed and is able to successfully maneuver the truck to the loading dock from any initial condition that was within a particular region. The region represented a group of non-trivial trajectories that the truck is to follow.

Global feedback methods of a hierarchical controller for real-time execution and replanning

Feedback methods for a hierarchical structure that is based on a physical decomposition of the system are considered. These methods consist of global and local feedback during real-time executions and they incorporate real-time replanning. Local feedback methods consist of utilizing all the available measurements. This availability of other measurements enables the implementation of feedback controllers at almost every node. Global feedback methods involve detection and propagation of changes and replanning due to these changes. Since these methods are different from the standard feedback methods, real-time execution and replanning processes are described in detail. One of the major advantages of the feedback methods described are their ability to replan as much as possible within a time limit.<<ETX>>

A new approach to secondary voltage control

Secondary voltage control has been applied in many countries, especially in Europe. The aim of the secondary voltage control is to minimize voltage deviations at load buses so as to improve the system voltage security. This paper addresses a new approach to show the improvement of system stability with the secondary voltage control. Two case studies based respectively on the 10-bus system and the New England 39-bus system are provided.