Research On Parameter Blind Estimation Of Frequency Hopping Signals

Frequency hopping(FH) communication has been widely used in military and civilian communications for its low probability of intercept, strong networking capability and good performance of jamming against. Researches on frequency hopping investigative technique is full of great reality significance. Blind parameter estimation of frequency hopping signals with only single receiving antenna is discussed in this paper, which is an important part of FH investigative technique. That is to say, it is neccesary to estimate the number of FH signals, the hop instants, the hop duration and the frequencies of each FH signal during every duration from the single receiving antenna without any prior information. The author’s major contributions are outlined as follows:1. The problem about estimating the parameters of multiple slow FH signals in synchronous orthogonal network is studied. An algorithm based on the combination of Short-time Fourier Transform(STFT) and Smoothed Pseudo Wigner-ville Distribution(SPWVD) is proposed. Apply the STFT to the signals, and then the rough ranges of the hopping times and the frequencies during each duration can be gotten. The accurate estimation of the hop instants is identified by employing SPWVD to the rough ranges, then the hop period is gotten. Experimental results show that this algorithm has better estimation performance and a certain noise resistance compared with the exisiting methods.2. The problem about estimating the parameters of multiple fast FH signals in synchronous orthogonal network has been studied. An improved algorithm based on the auto regressive and moving average(ARMA) model is presented. Some corrections of virtual points and leakage points of hop instants are added to the improved algorithm to estimate the parameters. Moreover, the method of frequency estimation is replaced with Fast Fourier Transform(FFT). With the improved algorithm, there is no need to know the number of FH signals of the sender in prior and the estimation error is smaller.3. For asynchronous non-orthogonal network, an new algorithm based on frequency join is put forward. Use STFT to the received FH signals after being sampled, then cluster the frequencies during each window. Calculate the start window number, end window number and the sustained windows of each frequency. Classify the frequencies according to the principles that the frequencies of the same FH signal can be stitched fore and aft and the hop duration of the same FH signal is the same. Then the frequencies of each FH signal during different period can be gotten. With the start time of each frequency, the hop duration and hop instants of each FH signal are obtained. Simulation results show that this algorithm can deal with the collision frequency and separate the FH signals of the same hop duration effectively. When the signal to noise ratio(SNR) is -6d B, the relative estimation errors of hop instants, hop durations and frequencies are both about 0.001.