Arq Analysis under Finite-state Hidden Markov Models Approved by Supervisory

ACKNOWLEDGEMENTS I would like to thank Dr. Aria Nosratinia for giving me an opportunity to do research as a graduate student with him. I express my gratitude to Dr. Nosratinia for his support, guidance and advice not only in my work but also in other aspects of my life. It is a wonderful experience to work with him. I would like to thank Dr. Fumagalli and Dr. Fonseka for serving on my supervisory committee. I also thank my colleagues at the Multimedia Communications Laboratory for their help and support. They are a special group of people-intelligent, enthusiastic and humorous. Thanks especially to Shahab for being very helpful in the beginning of my research. I would like to thank my family for their encouragement and support. I greatly thank my mother who inspired me to pursue Master degree and my father who gives me his full support for everything I do. In this work, we analyze two basic ARQ protocols, Go-Back-N (GBN) and Selective Repeat (SR). Previous analyses of ARQ protocol have concentrated on simple two-state Markov models. We solve a wider class of problems by characterizing both the forward and reverse channels by finite-state hidden Markov models. We show that a Hidden Markov Model (HMM) can describe a block fading channel. We then calculate the throughput of GBN with reliable feedback, as well as unreliable feedback using block transition probabilities. We also calculate the throughput and delay of SR ARQ under HMM. The moment generating function (MGF) technique is used to find throughput and delay. To calculate the MGF, we construct matrix signal flow graphs for the hidden Markov process. This procedure can be useful for a variety of other HMM problems and is therefore of interest by itself. Practical issues such as erasure errors and timeouts are included in our analyses, which are verified by extensive simulations. vi TABLE OF CONTENTS Acknowledgements v Abstract vi List of Figures ix Chapter 1. Introduction 1 1.