Single Carrier Underwater Acoustic Communications In Strong Interference Environment

Compared with the multi-carrier system,the single carrier(SC)system dose not have the problem of high peak to average power ratio(PAPR)and is insensitive to Doppler frequency offset.It is one of the optimal schemes to realize high-speed underwater acoustic communications(UAC).With the increasing of human activities in the marine and the expansion of information transmission scenarios,UAC system is facing performance degradation caused by strong interference.However,there are very few studies on SC-UAC system in strong interference environments.From the perspective of the system,this thesis divides the strong interference into two aspects: strong external interference and internal interference.Firstly,we study strong external interference,establish an external interference model in a SC system,and propose a corresponding iterative receiver structure.Then the external interference is sparsely constrained,and the channel equalization method in time-varying channel under sparse interference is further studied.Finally,the study of strong interference is extended to multi-input and multi-output(MIMO)system.MIMO channel estimation method under sparse external interference and MIMO internal interference suppression method with strong spatial correlation are studied.The effectiveness of the proposed method is verified by numerical simulation and experimental data.The first is to study the channel equalization in parameterized external interference.Based on Nyquist sampling theorem,the external interference is parameterized by a finite number of unknown parameters.Based on this model,an iterative receiver with interference cancellation is proposed: first,the external interference is detected and some parameters are estimated by Page Test,then the interference wake-up channel estimation is performed,then the interference is estimated,reconstructed and eliminated by using the above parameters,and finally the channel equalization and symbol decoding are carried out.The proposed receiver is evaluated by simulated partial-band and partial-block interference,impulsive noise and artificial sonar signals.Numerical simulation and experimental results show that the proposed equalization method can effectively overcome the influence of strong external interference,and has a greater performance improvement compared with the traditional decision feedback equalization(DFE).The second is to study the channel equalization in sparse external interference.Based on the parameterized model and sparsity constraint,a SC time-frequency domain equalization method in sparse external interference in time-varying channel is proposed.Firstly,for time-varying underwater acoustic channel,sparse channel tracking based on bi-directional Kalman filter is used to improve the channel estimation accuracy.Then,the initial symbol estimation is carried out by iterative receiver with parameterized interference cancellation,and the residual Doppler is compensated by group phase correction.Finally,the decoding performance of the receiver is further improved by the joint interference,channel and symbol estimation based on Sparse Bayesian learning.The proposed equalization method is evaluated by Gaussian mixture model(GMM)noise,S?S noise and Arctic impulsive noise.Numerical simulation and experimental results show that the proposed method has excellent decoding performance in sparse and strong external interference environment compared with traditional frequency domain equalization and iterative equalization based on parameterized model.The third is to study MIMO channel estimation in sparse external interference.Different from single-input and single output(SISO)or single-input and multi-output(SIMO)systems,the channels of different data streams in small-scale MIMO systems are usually spatially correlated.Based on this,a MIMO channel estimation method based on variational Bayesian learning is proposed.In the joint channel and interference estimation method based on Sparse Bayesian learning,the channel vector and interference vector are combined into one vector for overall estimation.There are two problems,and the first is that the large dimension of vector results in the increase of computation.The second is that it is assumed the sparsity of channel and external interference is similar in the joint estimation method which is not true in practice,therefore leads to the loss of performance.In the channel estimation based on variable Bayesian learning,the channel and interference vectors are estimated separately,and the spatial correlation between channels is considered,so as to reduce the complexity and improve the channel estimation accuracy.The proposed equalization method is evaluated by GMM noise,S?S noise and Arctic impulsive noise.Numerical simulation and experimental results show that the proposed channel estimation method has higher estimation accuracy than LS,OMP,SBL and ISBL channel estimation methods.The fourth is to study the MIMO internal interference suppression.The external interference suppression methods proposed in the previous chapters can be directly applied to MIMO systems.However,MIMO systems will face strong internal interference when the channels between different data streams are correlated.To solve this problem,a MIMO equalization method with strong spatial correlation is proposed.Based on the signal estimation theory,the symbol estimation under LMMSE criterion in single carrier MIMO systems is obtained,and then a time-domain interference rejection combination(IRC)method is discussed,which calculates the coefficients in frequency-domain and deconvolutes in time domain.This method has low computational complexity and can suppress co-channel interference theoretically.Using time-domain IRC as the preprocessing step of Turbo iterative receiver can further suppress the residual co-channel interference and inter symbol interference,therefore improve the decoding performance.Due to the interference preprocessing,the multi-channel received streams are combined into a single stream,which greatly reduces the computational complexity of subsequent Turbo equalization.In the experiments with two transmitters and four transmitters in Songhua Lake,the low bit error decoding is realized,which verifies the effectiveness of the proposed method.