Detection And Parameter Estimation Of Frequency Hopping Signal In The Background Of Impulse Noise

Frequency hopping communication is widely favored in various fields because of its anti-interference,low interception and difficulty in deciphering.The frequency hopping reconnaissance process mainly includes frequency hopping signal detection,parameters estimation and network stations sorting,which has been a research focus of non-cooperative frequency hopping communication technology.To complete the combat mission of frequency hopping communication,it is necessary to effectively intercept the frequency hopping communication signal.When the enemy's communication signal is unknown,the receiver first needs to detect the received signal to determine whether there is a frequency hopping signal.Under the premise of confirming the existence of the frequency hopping signal,the frequency hopping characteristic parameters are obtained through parameter estimation to detect the activity law of the frequency hopping network station,and the source signals belonging to the same network station are sorted according to the characteristic parameters,which can then be targeted for interference The enemy,protect the normal operation of your own communications.In the actual communication environment,the noise has the characteristics of random spikes,so the performance of detection and parameters estimation methods for frequency hopping signal based on Gaussian assumption is degraded.This paper studies the low signal-to-noise ratio Alpha stable distribution noise background frequency hopping signal detection and parameter estimation method,and the main work is summarized as follows.Firstly,the working principle of frequency hopping communication system and the related theory of Alpha stable distributed noise are briefly introduced.Secondly,a blind detection algorithm for frequency hopping signals based on time-frequency cancellation is proposed.On the basis of short-time Fourier transform,the different time-frequency characteristics of frequency-hopping signal and fixed-frequency signal are made full use of,and theoretically analyzed and the time-frequency cancellation ratio of frequency-hopping signal,fixed-frequency signal and Gaussian white noise are derive.According to the difference of time-frequency cancellation ratio,the ternary hypothesis test rule is obtained.The time-frequency cancellation ratio is obtained by using fractional low-order short-time Fourier transform instead of short-time Fourier transform to detect frequency hopping and fixed frequency signals under the condition of unknown prior information in Alpha stable noise background.The simulation results show that the detection performance of the algorithm is better than the power spectrum cancellation method,the anti-noise power uncertainty is stronger,and the implementation is simple.Then,a blind detection and parameter estimation algorithm for frequency hopping signals by using time-frequency sparsity is proposed.The time-frequency matrix is obtained by using fractional low-order short-time Fourier transform.The time-frequency matrix is averaged by frequency row to suppress fixed-frequency interference.Then time-frequency peak optimization,time-frequency matrix cleaning and enhancement processing are performed to reduce spectrum leakage and the influence of noise.And the frequency hopping signal is detected according to the continuity of the frequency hopping signal in the dwell time,and the frequency of the frequency hopping is estimated;and the hopping time and period are estimated according to the starting point and interval of the dwell time.The simulation results show that under low signal-to-noise ratio,the detection and parameter estimation performance of the algorithm is greatly improved,and it is better than existing algorithms.Finally,a mixed matrix estimation algorithm based on single source point detection is proposed.Fractional low-order short-time Fourier transform is used to obtain the time-frequency matrix,the carrier frequencies are estimated according to the jump processing,and the time-frequency matrix is divided according to the carrier frequencies.Then short-time Fourier transform and a single-source point detection criterion are used to further filtering the time-frequency points in each cluster.And then,the adaptive potential function is improved to estimate the angle paramaters of the corresponding time-frequency ratio,so as to estimate the mixed matrix estimation of each hop.The simulation results show that the algorithm effectively solves the problems of the mixed matrix estimation and accurate estimation of the number of network stations under Alpha stable distributed noise with low signal-to-noise ratio.