Research On Reconnaissance Algorithm For Multi-frequency Hopping Signals

Frequency hopping communication technology has excellent characteristics of strong anti-interference,and has been widely used in both military and civil fields in recent years.At the same time,frequency hopping communications reconnaissance also ushered in severe challenges.Since there are usually multi-frequency hopping signals in reality,conducting research on multi-frequency hopping signal reconnaissance techniques and seeking efficient algorithms for processing multi-frequency hopping signals has become one of the urgent and arduous tasks in the field of communication reconnaissance.In this paper,the research of multi-frequency hopping signal reconnaissance algorithm under single antenna reception is carried out,which mainly includes three links of multi-frequency hopping signal extraction,parameter estimation and sorting.In the process of extracting multi-frequency hopping signals,firstly,an improved algorithm based on the combined time-frequency distribution of the spectral graph and the smooth pseudo-Wignaville is proposed.The performance of the algorithm is significantly improved without increasing the amount of calculation.The simulation results show that the improved spectrogram and smoothed pseudo-Wagnerville(SP & SPWVD)combined time-frequency distribution algorithm has a clearer time-frequency graph and lower quantitative information entropy than other combined time-frequency distributions.Secondly,an algorithm based on morphological filtering is used to denoise the time-frequency map.When removing fixed-frequency interference,an incomplete defect is eliminated for the existing morphological corrosion subtraction method,and a peak area based on connected area is proposed.law.The simulation results show that the denoising effect of the connected area area peak method is better.In the parameter estimation of multi-frequency hopping signals,a parameter estimation algorithm based on the connectivity of time-frequency graphs is proposed.When estimating the number of frequency-hopping signal sources,a method based on statistics of connected regions is given,which is similar to the existing methods.It is simpler and more efficient.When estimating the center frequency of each hop signal,a more accurate estimate can be obtained by extracting the position of the frequency point where the centroid of the connected area is located.Simulation results show that the overall performance of the proposed algorithm when applied to slow frequency-hopping systems is better than that of fast frequency-hopping systems.When the signal-to-noise ratio of the fast-hopping system is relatively low,the estimation error increases by an order of magnitude,and the parameter estimation performance appears larger.deterioration.Finally,the simulation compares the performance of parameter estimation based on the two combined time-frequency distribution algorithms before and after improvement,and verifies the effect of the time-frequency graph results on the proposed algorithm and the effectiveness of the improved combined time-frequency distribution algorithm.In the multi-frequency hopping signal sorting link,for the asynchronous non-orthogonal network with different hopping periods,a multi-frequency hopping signal sorting algorithm based on K-Means clustering of period parameters is proposed.By using the number of frequency-hopping signal sources as the number of clusters,the process of calculating the loss function to find the K value is omitted.The simulation results show that the algorithm has a very high sorting accuracy when the mean square error of the hopping period is small.For the same frequency hopping period,the initial sorting is based on the time continuous characteristics of the same frequency hopping signal,and then the second sorting is based on the similar energy characteristics of the same frequency hopping signal.Finally,when the frequency collision between the fixed frequency interference and the frequency hopping signal occurs,a time-based loss-hop detection algorithm and an energy-based loss-hop recovery algorithm are proposed.The effectiveness of the proposed algorithm is verified by simulating different situations between jammed hops.