Research On The Recognition Algorithm Of FH Signals

As a fast developed communication style, the frequency-hopping communication, for its excellent anti-interference ability, has been widely applied both in military communication and civil uses. Especially in the field of electronic counter, in order to occupy the favorable situation, it is necessary not only to transfer information secretly, but also obtain enemy's information as far as possible. Thus, the detection and parameter estimation for the frequency-hopping signal become an important research direction in communication field recently.As a typical non-stationary signal, frequency-hopping signal can not be analyzed by traditional methods. In this paper, a kind of representation named time-frequency representation is adopted to analyze the frequency-hopping signals, which reflects both the frequency contents and the contents'changing law along with the time. As a matter of fact, WVD is one of the most popular used bi-linear time-frequency representations. Whereas, WVD has severe cross-term interferences for the multi-frequency components signals. Actually, they can be reduced by a two-dimensional low pass filtering of WVD. Unfortunately, this smoothing also produces a less accurate time-frequency localization of the signal components. To overcome these shortcomings, in this paper, a reassignment method is adopted, which is called RSPWVD, so that there are both less cross-terms and a good concentration of the signal components. So far, most of the simulation, which is to estimate the parameters of the frequency-hopping signals, is realized by the way of software. And the information of the sources and channel is assumed. To solving these problems, in this paper, the simulation environment of wireless radio platform which is based on the GNU Radio and USRP, is adopted to realize the frequency-hopping signals'transition, receive and parameter estimation.Targeting at the blind parameters estimation for frequency-hopping signals, this paper introduces three kinds of popular used time-frequency representations and estimates the parameters of the frequency-hopping signals with one of them. Meanwhile, in order to improve the performance of the algorithm, in this paper, the reassignment method is adopted. Meanwhile, the estimation results have been obtained based on the GNU Radio and USRP and the performance of this kind of algorithm has been shown through analyzing the results.