On The Performance Of Space-Time Coding Over Fading Channels In Non-Gaussian Noise

The goal of future wireless communication systems is to provide higher data rates and better quality services. One major technological breakthrough that will make this increase in data rates possible is to use multiple-input multiple-output (MIMO) system. Being a key technology of future wireless communication systems, MIMO technology can significantly improve capacity over fading channels, error performance and spectral efficiencies without any loss in bandwidth and transmitting power. Space-time coding (STC) is a set of practical signal design techniques aimed at approaching the information theoretic capacity limit of MIMO channels, which is used for multi-antenna systems based on MIMO technology. Since the fundamentals of space-time coding was established, it has been paid great attention to and studied deeply. It is noted that most researches are performed on the assumption that the communication systems are working in Gaussian white noise. However, there are many phenomena in communications that contribute in producing noise that has a non-Gaussian distribution, that is to say, non-Gaussian noise is prevalent in wireless environments and often impulsive in many applications. Therefore, a more realistic noise environment would be non-Gaussian. Different from previous researches, this dissertation is concentrated on the performance analysis of STC over fading channels in non-Gaussian noise environments. The main contents and contributions of this dissertation are summarized as follows:1. Based on the moment generating function (MGF) approach, a method of performance analysis of orthogonal space-time block coding (OSTBC) over fading channels is proposed in this dissertation. Closed-form expressions for the exact symbol error rate (SER) of OSTBC with M-PSK and M-QAM modulations over independent identically distributed (i.i.d.) fading channels such as Rayleigh, Rician and Nakagami-m as well as correlated Nakagami-m are derived in terms of the MGF of instantaneous SNR of SISO channels. The performance analysises of OSTBC over various typical fading channels are performed from the aspects of coding schemes, spectral efficiencies, modulations and channel correlation. Also some new observations on the performance results of different code matrices are made.2. Propose a method of performance analysis of OSTBC over fading channels in a mixture of impulsive noise and Gaussian noise environments. Anε-mixture Gaussian noise model for the noise distribution is adopted and the channels are modeled as Rayleigh, Rician and Nakagami-m flat fading, respectively. Based on an equivalent scalar additive mixture noise channel, the approximated MISO model of OSTBC in mixture noise can be constructed to obtain the MGF and probability density function (pdf) of the instantaneous SNR. The closed-form expressions for the average SER of OSTBC over fading channels under the impulsive model are derived by means of the MGF and directly calculating the integral of the pdf. Furthermore, the SER performances of the OSTBC are evaluated for different code matrices with MPSK and MQAM modulations. Simulation results show fine agreement with the analyses.3. The performances of OSTBC in Middleton Class A and a-stable noise are investigated in this dissertation. Firstly, as to the statistical-physical Middleton's Class A noise model, by transforming non-Gaussian distribution into conditioned Gaussian distribution, we can get the joint pdf of the noise signal in the OSTBC system model under the assumption of dependent in space and independent in time. Then the upper bound on the average pairwise error probability (PEP) is derived for the maximum likelihood (ML) receiver in Middleton's Class A noise. Secondly, through introducing the concept of equivalent dispersion parameter, geometric power and geometric SNR, we provide a mathematically and conceptually characterization of the relative strength between the information-bearing signal and channel noise with infinite variance. Also, Monte Carlo simulations with respect to above two scenarios are performed with results and analysis conclusions given.4. Propose a method of performance analysis of space-time trellis (STTC) over fading channels in a mixture noise environment. Utilizing the MGF-based approach, the exact expressions for PEP of STTC over Rician and Nakagami-m fading channels are derived in terms of the single variable integral form. Furthermore, based on the relationship between the SNR for Gaussian and that for mixture Gaussian, the approximated PEP in mixture noise is derived. With this PEP, an analytical estimate for bit error rate (BER) is evaluated, which is over slow and fast fading channels, respectively.5. Modeling and simulation of fading channels and non-Gaussian noise signal are investigated. The methods based on the inverse transform approach of generation of random variable sequences obeying Rician as well as Nakagami-m distributions are presented, respectively. These random variable sequences generated in this dissertation are generalized and can be used in the simulation of wireless communications widely. The scheme of using histogram method is also proposed to generating non-Gaussian random variables, of which the effectiveness is verified by the simulation experiments.