Adaptive wavelet network PID controller for improving the performance of active queue management in the TCP network

Dissertant

Hasan, Jalil Hasan

Thesis advisor

al-Qarawi, Salih
Doraiee, Faruq

University

University of Technology

Faculty

-

Department

Department of Control and Systems Engineering

University Country

Iraq

Degree

Ph.D.

Degree Date

2013

English Abstract

In this dissertation, an adaptive wavelet network Proportional Intégrai Derivative (AWNPID) controller is developed for Active Queue Management (AQM) in an end-to-end TCP network It is based on wavelet network with an Infinite Impulse Response (IIR) recurrent structure.

The IIR is combined by cascading the network to provide double local structure resulting in improving speed of learning.

AWNPID controller is designed to overcome the PI and PID problems.

It adaptively controls the dropping probability of packets by training the wavelet network parameters (weight, dilation, translation, and feed- forward and feedback IIR coefficients), revising these parameters online by Least Mean Square (LMS) learning rule, and then tuning the PID controller coefficients according to the wavelet network output.

The analytical results for linearized TCP/AQM model are presented using MATLAB version 7.0.

NS2-simulator (version 2.27) is also used to support the proposed design.

Main Subjects

Information Technology and Computer Science

Topics

• Computing

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-417778