Symbol synchronization and dynamic channel estimation in Multiple Input Multiple Output systems using statistical analysis of eye diagrams

The research work presented in this document is the continuation of the project to develop a Multiple Input Multiple Output (MIMO) system using band-limited signals in strong multipath environments. The modulation format used is Gaussian Frequency Shift Keying (GFSK). Band-limiting introduces controlled Inter-Symbol Interference (ISI), which along with a multipath environment degrades the features of the received signal waveform that are vital to the performance of standard symbol synchronization techniques. A low complexity algorithm was deemed necessary to counter this problem so that the benefits of conserving bandwidth would not be obviated by an expensive and elaborate receiver architecture. Such an algorithm, based on the standard deviation of the eye-diagram of the received signals was developed. The performance of the algorithm was analyzed under different channel and system conditions. The algorithm could not only track variable baud rates but also provide optimum symbol sampling points, even for highly obscure eye-diagrams. Another aspect of the project was tackled using the same principle of standard deviated of eye-diagrams: Dynamic Channel Estimation. MIMO systems require recurrent transmission of training symbols to maintain updated channel estimates. The rate of this transmission was determined by an algorithm that correlated statistical behavior of received signals with bit error rates --Abstract, page iv