| Summary: | 博士 === 國立臺灣海洋大學 === 電機工程學系 === 99 === In this dissertation, we propose a physical layer infrastructure of a robust underwater acoustic communication system and perform theoretical analyses for the system. We first propose a design rule of an SER optimized APSK modulation. By adopting a multi-ring constellation architecture, our design rule can unify the Euclidean distance between any two nearby symbols within the constellation and achieve the desired SER with minimum symbol energy. Our computer simulations prove that the SERs of the proposed APSK constellation in both the AWGN and Rayleigh fading channels match the results predicated by theoretical analyses. In addition, by setting the phase noise free space properly, the optimal design rule can maintain the minimum Euclidean distance on intra-ring symbols to avoid the increase of SER caused by ICI. Analyses and simulations reveal that our optimized APSK modulation has better SER than QAM when ICI exists.
To reduce PN code acquisition time for underwater acoustic communication systems based on spread spectrum technology, we further propose a new Coarse/Fine PN code acquisition scheme. By using short-length correlator sets, we transform the received spreading signal into a sequence of normalized correlation values. The trained BPNN can identify the sequence quickly and find out the phase of received spreading signal. In addition, the local PN code generator uses the estimated PN code phase to generate verification sequence and calculate the correlation between the subsequent received spreading signal and the check sequence. By evaluating the correlation value, the system can tell if the estimated PN code phase is correct. When compared to traditional PN code acquisition systems, our proposed system only needs brief sections of received spreading signal to identify the phase of PN code. For bandwidth-limited underwater acoustic communication systems, our PN code acquisition system can do fast acquisition and reduce the preambles for synchronization. Simulation results have confirmed that our proposed system can quickly acquire the phase of received spreading signal with random initial phase as anticipated by our theoretical analyses.
|
|---|
