수중 음향 통신에서 추정 BER을 이용한 다중밴드 PSK 전송 기법 연구

Abstract

Due to the increasing consideration of ocean environment and resources, there is a growing need of reliable underwater acoustic communication. In the underwater environment, the multi-band technique has been widely researched because it can overcome selective frequency fading which is caused by multi-path propagation by allocating the same data to different frequency bands. Thus, the multi-band technique is effective in terms of performance and throughput efficiency. However, in some underwater environments, a particular inferior band can affect the output from the entire bands. To overcome such problems, a receiving end that allocates weights according to performance analyzation of each band can be applied. However, those conventional methods have limits in that they can not analyze coded bit area which have the actual information but only the preamble area. So in this thesis, an algorithm to set the weighting value using an estimation BER (Bit Error Rate) is proposed. Estimation BER is the method of analyzing reliability of received data based on the performance difference between demodulated and decoded data. the parameters which affect the performance of underwater acoustic communication are figured out and a method to set the optimal parameter values is suggested. Through the simulations, it is figured out that the performance of multi-band is improved when the proposed threshold algorithm is applied. To analyze the efficiency of multi-band transmission scheme with estimation BER in underwater communications, a short distance lake experiment and a long distance sea experiment was conducted by employing phase shift keying modulation and turbo codes with 1/3 of coding rate. Through the experiments, the performance of conventional methods and the suggested method was compared. When the conventional method couldn't correct the errors, suggested method using estimation BER corrected the errors perfectly. Also, it is confirmed that the amount of Doppler shift can be different on each data and preamble field. Therefore, the proposed algorithm is well-operated in the environment of time-varying underwater communication channel.