Signal Recognition And Parameter Estimation For Multi-carrier Systems In Shortwave Communications

Orthogonal Frequency Division Multiplexing (OFDM) and Multi-tone Parallel modulation are tow kinds of multicarrier techniques having the advantages of robustness against multi-path fading, high bandwidth efficiency. And due to their low complexity, they have been widely used in shortwave communications.This dissertation is devoted to the study of the recognition and parameter estimation of multi-carrier signals. The work finished in this paper is a part of a large scale army project of research undertaken by the lab the author works with. The main work and innovative algorithms proposed are summarized as follows:1. Based on the correlation of cyclic prefix of received signals, an algorithm is proposed for blind recognition of OFDM signals. This algorithm can distinguish OFDM signals from the others, and it can be widely used. Simulation results show that it has higher accuracy in recognition with little data.2. The correlation of the cyclic prefix can be expanded to two or several signals. The parameter estimations, such as estimation for the length of symbol, cyclic prefix length and symbol rate, is obtained by searching the peak value of the combined correlation function. Experimental results show that the estimator can provide more accurate estimation for practical signals.3. Estimation for the number of practical sub-carriers is a difficult point in OFDM parameter estimations. This paper investigates the high-order cyclic cumulants to estimate the number of sub-carriers of OFDM signals in frequency-selective fading channel, which is proved by theoretical analysis. Simulation results show that this method can obtain good performance even under Rayleigh fading channels, which make it suitable for the implementation in shortwave signals.4. In order to distinguish the multi-tone parallel signals from single-carrier ones, this paper uses an improved method to search the peek value of the spectrum. Targeting at the mistaken recognition of the MFSK signals as multi-tone parallel signals, further recognition is obtained by wavelet transform.5. Particular template matching method is designed to identify the 12 multi-tone parallel signal. Some other multi-tone parallel signals can be identified by changing the identification parameters. Experimental results from the tests of practical signals show this algorithm can obtain more efficient matching.