Sensitivity of OFDM Systems to Synchronization Errors and Spatial Diversity
Abstract
In this dissertation, the problem of synchronization for OFDM-based wireless communication systems is studied. In the first part of this dissertation, the sensitivity of both single input single output (SISO) OFDM and multiple input multiple output (MIMO) OFDM receivers to carrier and timing synchronization errors are analyzed. Analytical expressions and numerical results for the power of inter-carrier interference (ICI) are presented. It is shown that the OFDM-based receivers are quite sensitive to residual synchronization errors. In wide-sense stationary uncorrelated scattering (WSSUS) frequency-selective fading channels, the sampling clock timing offset results in rotation of the subcarrier constellation, while carrier frequency offsets and phase jitter cause inter-carrier interference. The overall system performance in terms of symbol error rate is limited by the inter-carrier interference. For a reliable information reception, compensatory measures must be taken.
The second part of this dissertation deals with the impact of spatial diversity (usage of multiple transmit/receive antennas) on synchronization. It is found that with multiple transmit and receive antennas, MIMO-OFDM systems can take advantage of the spatial diversity to combat carrier and timing synchronization imperfections. Diversity can favorably improve the synchronization performance. Data-aided and non-data-aided maximum likelihood symbol timing estimators for MIMO-OFDM systems are introduced. Computer simulations show that, by exploiting the spatial diversity, synchronization performance of MIMO-OFDM systems in terms of mean squared error (MSE) of residual timing offset becomes significantly more reliable when compared to conventional SISO OFDM systems. Therefore, spatial diversity is a useful technique to be exploited in the deployment of MIMO-OFDM communication systems.
In MIMO systems with synchronization sequences, timing synchronization is treated as a multiple hypotheses testing problem. Generalized likelihood ratio test (GLRT) statistics are developed for MIMO systems in frequency flat channels and MIMO-OFDM systems in frequency selective fading environments. The asymptotic performance of the GLRT without nuisance parameters is carried out. It is shown that the asymptotic performance of the GLRT can serve as an upper bound for the detection probability in the presence of a limited number of observations as well as a benchmark for comparing the performances of different timing synchronizers.