Research On Narrow-Band Interference Suppression Technology In DSSS System

Spread spectrum communication technology is widely used in military and civil communications due to its advantages of anti-jamming,anti-multipath,anti-fading and strong concealment.Direct Sequence Spread Spectrum(DSSS)technology directly expands the spectrum of the signal by multiplying the spread spectrum(SS)code with signal sequence.The length of the spread spectrum code determines the anti-jamming tolerance of the DSSS system.However,the SS code can not be infinitely long,so the anti-jamming tolerance has an upper bound.When the interference power is too large,beyond the bearing range of DSSS system,it will cause irreparable damage to the communication system.Narrow-band interference(NBI)is very destructive because of its power concentration and certain bandwidth.In order to reliably extract useful information from received signals with the existence of strong NBI,high performance algorithms are required to suppress interference.According to the problem of energy loss and limited application range in frequency domain interference suppression algorithms,starting from the existing frequency domain suppression techniques,threshold sideband algorithm is studied by combining the advantages of two frequency domain algorithms.Aiming at the problem of non-linear representation in time domain,pipelined recurrent neural network(PRNN)algorithm based on SS code is studied by combining spread spectrum code information with PRNN.The specific work includes the following:1.NBI suppression technology in time domain is studied.Emphasis is laid on time-domain linear and non-linear predictive filtering.The common NBI modes are studied,the mathematical expressions of interference are given and simulated.Finally,two kinds of time domain prediction techniques are analyzed experimentally.Nonlinear filtering has better BER performance.2.In view of the convergence problem of time-domain interference suppression algorithm and the shortcomings of high computational complexity,frequency-domain NBI suppression technology is studied.Firstly,the windowing technology before FFT transform is introduced,and the comparison between different window functions is given.The overlapping windowing technology is studied.Blackman-Harris window has excellent sidelobe suppression performance.Then the frequency domain adaptive filtering algorithm,frequency domain threshold algorithm and sideband correlation permutation algorithm are studied.Aiming at the problem that the frequency domain threshold algorithm loses a lot of signal energy and that the sideband correlation permutation algorithm can not deal with the interference at the symmetry of the sideband,a sideband threshold algorithm is proposed.The algorithm adopts corresponding processing methods for different spectral lines in frequency domain.The simulation results show that the proposed sideband threshold algorithm can achieve better performance than other frequency domain algorithms.3.Aiming at the non-linearity of prediction error in time-domain technology,neural network technology is introduced.The recursive neural network(RNN)is used to replace the weight connection in adaptive filtering,and the prediction error is updated iteratively by despreadrespread using the information of SS codes,and the real-time weight adjustment of RNN is realized by RTRL algorithm.The simulation results show that the recursive neural network technology based on SS codes has strong nonlinearity,can accurately predict NBI,and the suppression effect is obvious.