Research On Interference Detection And Blind Source Seperation In Frequency Hopping Comunication

With the information-based war intensified,signal processing has been more and more widely used in information confrontation.Frequency hopping communication system usually use large bandwidth to increase the anti-interference ability,but is limited.When the electromagnetic environment of the communication channel is very complex,source signal is affected by various types of interference and receiving signal is unsatisfactory.In this paper,we used the advantages of blind signal separation and combined with interference suppression technology,improved the system of anti-interference ability.The main contents are as follows:In the first part,the interference detection of frequency hopping based on time-frequency analysis is studied.Several kinds of interference signals including sweep frequency interference,multi-tone interference,periodic pulse interference,random inter-ference combined into an adaptive form,establish interference library,and transform to the frequency domain.According to the features of each signal in the time-frequency domain,we can filter the time-frequency matrix,contrast to the signal of interference library,and record time-frequency point of various jamming signals.In order to reduce the impact on the back of interference signal detection,we need to remove time-frequency point.However, due to the restriction of cross-over point,we need to process newly. Finally, we analyse the question of interference detection in the frequency hopping communication signal, in order to reduce the damage to the frequency hopping signal, the power recovery method of frequency hopping signal is increased.In the second part,we study the blind signal separation algorithm.The time-frequency pattern of frequency hopping signal is obviously showed the characteristic of "jump", which belongs to the non-steady signal.Considering three signals and two observation signals circumstances,the waveform in the time domain is overlapped,and can not be separated.The time-frequency characteristics of frequency hopping signal can be observed clearly when the observation signal is transformed to the frequency domain by using the short time Fourier transform.Due to the influence of noise and cross terms,the frequency hopping pattern showing a lot of noise.In this paper,in the original algorithm is improved,we first screen time-frequency point,detect single source time-frequency point,and set threshold again.The symbol vector is judged according to the maximum value,and the algorithm can adaptively select the column of time-frequency matrix.Finally,this chapter anlyse and compare with time frequency masking algorithm and higher order accumulated amo unt of blind signal separation algorithm.In the third part,we study the application of interference detection in blind signal separation algorithm.In the blind signal separation,the receiver receives many source signals,and increases the impact of interference signal,the time-frequency point is seriously affected.After testing,the operand of detection and identification process is enormous in the third chapter,so in order to depress the amount of computation and reduce the operand of cross term signal power recovery process,thid chapter study narrowband interference suppression algorithms based on wavelet packet and non-extraction wavelet packet.The simulation experiment shows that the undecimated wavelet packet algorithm of the interference suppression ability is excellent.Finally,the paper combined with frequency domain interference suppression algorithm and time-frequency domain interference detection and recognition algorithm.The interference signal is suppressed and the frequency hopping signal is reconstructed using the algorithm of the previous chapter.According to the correlation coefficients matrix,this algorithm can recover source signal,but the cumulative error is relatively large.