Reliable Transmission Techniques For Optical Wireless Communication Across Dynamic Water-Air Surface

Since optical spectrum can achieve communication in both air and water with sophisticated transceiver devices,it has great potential application in communication across water-air surface.Due to the effect of water flow,underwater communication nodes inevitably suffer dynamic drift and rotation.Moreover,the fluctuation of waterair interface cause random light refraction.As a result,the received signal strength varies significantly during a short time period.This thesis focuses on the large dynamic characteristics of wireless optical communication across water-air surface,and we aim to realize reliable transmission from the perspective of channel coding.Based on the designed channel coding scheme,we also give the feedback scheduling strategy under different fading characteristics.In this case,the communication system can adaptively adjust the code rate according to different channel environments to obtain maximum throughput.The main research contents and contributions of this thesis are as follows:The design of channel coding schemes for dynamic water-air channel.In this thesis,the characteristic of water-air channel is summarized as random extremely weak link with noisy,which can be modeled as erasure and noisy.Based on the erasure and noisy characteristics of water-air channel,we propose two channel coding schemes,namely the concatenated RS-LDPC code and Raptor code with an interleaver.The coding structure and overhead calculation of two schemes is given respectively,and an algorithm of channel estimation is proposed for decoding.Then,we build a water-air visible light communication offline system with a green LED as transmitter.A wave generator is used to emulate the wavy water surface with different wave amplitudes and frequencies by changing the strength and frequency of the push board,respectively.The thesis firstly introduce the method of link gain extraction in our system,and explore the coherence time of water-air channel.Then,we experimentally demonstrate the BER and FER performances of the two schemes under different water depths and different wave environments.Finally,the comparison of the two codes with the same overhead and approximately the same code length is given.The design of feedback scheduling schemes for dynamic water-air channel.Based on the RS-LDPC scheme,we design feedback scheduling schemes to choose the appropriate code rates under different channel environments,aiming to obtain maximum throughput.Firstly,an algorithm is proposed to generate decoding data set of RS-LDPC code under different SNR and wave environments based on finite blocklength information theory.For the rapidly time-varying characteristic of water-air channel,we proposed the scheduling method based on state transition probability.The proposed method is compared with some other channel prediction algorithms based on the instantaneous channel state information.The results show that the instantaneous channel state information under water-air channel is hard to track,and the prediction algorithms based on the instantaneous channel state information can not bring significant gain on throughput performance.Furthermore,considering the statistic characteristic of water-air channel,we transmit with long code block and slow feedback frequency.A mapping table of channel state and code rate is constructed,and the criteria of our mapping table is to get maximum throughput.Compared with algorithms based on the instantaneous channel state information,the feedback scheduling schemes based on the statistic characteristic of water-air channel have a 20%-30%improvement on throughput performance.