Web-based radio fading channel modeling, estimation and identification.

Abstract

This thesis focuses on the practical implementation issues for modeling, simulation, estimation and system identification of wireless fading channels. The state space models are introduced to describe fading channels including flat fading and frequency-selective fading channels. Without measurement data, simulation and estimation of fading channels are related to the motion of the transmitter or the receiver. The state space realizations and the Kalman filtering are derived for Rayleight fading, Ricean fading and frequency-selective fading channels. The inphase, quadrature and envelope components of the received signals are simulated and estimated through the state space realizations and the Kalman filtering. These model parameters are defined by the transfer function of a given Doppler power spectral density (DPSD) associated to several physical factors such as carried frequency, mobile moving speed, signal-to-noise ratio (SNR), an angle and an interval time. With measurement data, system identification is based on the Expectation-Maximization (EM) algorithm together with the Kalman filtering. These system parameters are computed recursively. Various identified results illustrate the processes of system identification for flat fading and frequency-selective fading channels based on the measurement data provided by Communications Research Center Canada (CRC). Measurement data could also be input by users. Finally, a web-based wireless fading channel simulation and estimation system is analyzed, designed and implemented using UML(TM) techniques and Java(TM) programming language. The web-based system can provide friendly interfaces to complete modeling, simulation, estimation and system identification of fading channels. The Web-based simulation and estimation system can be run on the following site: http://www.site.uottawa.ca/∼jzhan037/onlineSystem.html or http://www.site.uottawa.ca/∼chadcha/onlineSystem.html .