Outage Probability Analysis of Correlated Sources Transmission over Fading Channels with Line-of-sight Component

In this thesis, the outage probabilities of correlated sources transmission systems are analyzed. The primary goal of this work is to derive theoretical limit for correlated sources transmission which utilize the source correlation, and to establish theoretical bases for wireless cooperative communications system design utilizing the correlated sources transmission concept. We focus on the problems that correlated sources are transmitted via channels having different statistical properties, and analyze how significant influent the differences of the statistical channel characteristics makes on the system outage. First of all, in this thesis, the outage probability of a system having two correlated sources transmitted to common destination through Rayleigh and Rician fading channels is derived. The outage probability can be expressed by double integrals with respect to the probability density functions (PDF) of the instantaneous signal-to-noise power ratios (SNRs) of each channel, where the range of the integration is determined by the Slepian-Wolf theorem. This work identify the effect of the Rician factor K on the outage probability, where the K factor denotes the average power ratio of the line-of-sight (LOS) component power-to-non-line-of-sight (NLOS) components. To verify the consistency between the theoretical and practical results, we apply the technique to a Slepian-Wolf correlated sources transmission system where bit-interleaved coded modulation with iterative detection (BICM-ID) scheme is used. Performance comparison between the theoretical outage and the frame-error-rate (FER) shows that FER curves exhibit the same tendency as the theoretical results. Moreover, we investigate the optimal power allocation for minimizing the outage probability with the condition that the total transmit power of the two sources is kept constant. The analytical results show that, so far as two sources are not fully correlated, lower outage probability can not always be achieved by increasing the transmit power ratio allocated to the source which transmits the signal via the channel have LOS component. We then investigate the outage probability of a system having two correlated sources transmitted to common destination via Rayleigh and Nakagami-m fading channels, where the Nakagami-m fading channel model is known to more accurately represent the distribution of fading variations of the channel having LOS component than Rician fading channel model. The Nakagami-m distribution well represents the channel variations because it is derived from the measurement data gathered in real fields. The same analytical techniques as the one that are used in the Rician fading case is applied to calculate the outage probability. The most significant contribution of this work is the derivation of a close-form expression for outage probability in several extreme cases. Furthermore, the asymptotic decay of the outage curves are derived theoretically. In the main body of this thesis, we analyze the outage probability of a Slepian-Wolf correlated sources transmission system over Rayleigh and Rician fading channels, as well as over Rayleigh and Nakagami-m fading channels, respectively. Rician fading model can well be approximated by the Nakagami-m fading model by adjusting the factorK in Rician fading model and the factor m in Nakagami-m fading model. Hence, it is quite meaningful to identify the impact difference on outage performance. We investigate Kullback-Leibler distance (KLD), which is a measure of the difference between two probability distributions, between the Rician and Nakagami-m distributions. We then evaluated the impact of the KLD on the outage performance.