Research On Key Technology Of Underwater Acoustic Communication Based On OFDM

With the development and mature of the wireless ocean career, the underwater acoustic (UWA) communication technology has got people's widespread attention. Because the ocean channel is such a complex environment that it varies in space, time and frequency, and have strong mutipath, reliable and high data rate UWA communication becomes a hotspot in UWA research field at present. In the OFDM system, high-speed data stream through bunch and merge transform, makes data symbol duration increase relatively in each sub-carrier wave, reduces inter-symbol interference from mutipath effectively, each subcarrier transmit data in parallel, which has high utilization of frequency spectrum and posses the ability of resisting frequency selective fade. In this thesis, OFDM is applied in UWA communication, and the UWA communication system based on OFDM is proposed. some reserches are follow:Firstly, the bandwidth of the UWA channel is limited, aiming at this characteristic, this paper applies TCM technology in UWA signal modulation. Introducing the theory of TCM, TCM-8PSK and their application in the UWA communication based on OFDM. By the experiments it is proved that this modulation mode can solve the data transmission efficiency problem and reliability problem with good performance.Secondly, this article analyses respective characteristics and applies occasion of block pilot and comb pilot, by experiments the performance of them under differents marine environment are compared. This article also carries out the LS estimator, MMSE estimator and advised estimation method, then study the performance of these estimation algorithms.Finally, based on the sparseness of UWA channel, this article applies Compressed Sensing Theory to the channel estimation, proposes a channel estimation algorithm based on approximate l₀ norm. The algorithm obtains the minimum l0 norm with the optimal solution of approximate function, thus obtains the sparse channel response. Simulation results show: under the same conditions, when signal-to-noise ratio is the same, compared with traditional estimation algorithms the MSE of this algorithm is small and have low calculation burden.