Research On Key Technologies For Communication Signal Receiver Under Non-gaussian Noise

In the field of communication and other signal processing,noise is often modeled as Gaussian distribution based on the physical mechanism of noise generation and the central limit theorem.However,in many natural and man-made noise environments,large-amplitude spikes frequently occur in the observed noise samples so that the noise is impulsive,which is quite different from the Gaussian noise model.These scenarios include: atmospheric noise in very low frequency/low frequency(VLF/LF)communication systems,acoustic noise in underwater acoustic communication channels,telephone line noise,network interference in wireless communication networks without power control,etc.In such impulsive noises,the performance of the signal processing algorithm,which is developed based on Gaussian noise model,will significantly deteriorate due to the mismatch between the noise model and the practical noise.Therefore,many researchers have conducted extensive and in-depth researches on the impulsive noise,and ? stable distribution is commonly adopted for impulsive noise model.The ? stable distribution can be derived based on the generalized central limit theorem and has been widely used in many research fields.This disseration focuses on the receiving technologies of communication signals under ? stable distributed noise.The main research contents and contributions can be summarized as the following four aspects:First,the statistical properties of ? stable distributed noise are discussed,and a robust noise model parameter estimation method is proposed.The statistical theory of ? stable distribution is discribed,and the impulsive noise is modeled as symmetric symmetrical ?-stable(S?S)distribution in this disserataion based on the existing works.Then,when both the transmitted data and noise are included in the observation data,a novel robust noise parameters estimation algorithm is proposed based on Finite Impulse Response(FIR)filtering.Comparing with the conventional parameter estimation algorithm,numerical simulation results show that the proposed parameter estimation algorithm can achieve smaller mean square error(MSE).Second,the synchronization technologies of communication receiver is addressed under S?S noise.For both linear modulation(LM)and continuous phase modulation(CPM)signals,the corresponding Cramer Rao Lower Bound(CRLB)is derived,and an optimal synchronized sequence is then given by minimizing the CRLB.The optimality of the proposed synchronization sequence is verified by comparing with the commonly adopted pseudo-random synchronization sequence.Furthermore,three synchronization algorithms are proposed,which are the maximum likelihood measure based gradient method,the robust measure based global optimization method,and the fractional lower order moment measure based search method.The performances of these algorithms are compared through numerical simulation.The results show that the first two algorithms can achieve performances close to the CRLB at the cost of high complexities,and the performance of the last algorithm has a gap to the CRLB but it has lower complexity.Third,the communication signal demodulation technologies for both continuous phase frequency shift keying(CPFSK)and LM signals are investigated under S?S noise.For CPFSK signal,the S?S process with independent and identically distributed samples is adopted to model the white impulsive noise,and the corresponding coherent and non-coherent demodulation algorithms are presented.The theoretical SER performance of the proposed algorithm is further analyzed.Meanwhile,the ?-sub-Gaussian process with correlated samples is used to model the burst impulsive noise,and the coherent and non-coherent demodulation algorithms of CPFSK signal are also proposed.To demodulate the LM signal with shaping pulse occupying multiple symbol periods under S?S noise,a multi-symbol-assisted signal demodulation algorithm is proposed,and its theoretical SER performance is analyzed as well.The SER performances of the proposed demodulation algorithms are evaluated by numerical simulations.The results show that the theoretical SERs match the simulation results well.Fourth,the communication signal preprocessing technologies are studied under S?S noise,so that the signal receiving technology designed for Gaussian noise can still be used.Both memory and memoryless preprocessing methods are considered.For the preprocessing technology with memory,the traditional myriad filtering method is improved to adapt the band-pass receiving system.For the memoryless preprocessing technology,the optimal preprocessing method under S?S noise is discussed.Simulation results show that the proposed optimal memoryless preprocessing method can achieve nearly the same performance as the myriad filtering based preprocessing method.Note that the complexity of the former is much lower than that of the latter.Moreover,the Clipper preprocessing method is addressed,which is commonly used in practice.Through theoretical derivation,the calculation method of the optimal threshold of Clipper preprocessing is given.Numerical simulation is also performed to verify the optimiality of the proposed optimal threshold.Finally,the above mentioned theoretical results and algorithms can provide guidlines and potential schemes for pracitcal communication system design under S?S noise.