Research On The Algorithm Of Shortwave Communications Signals Modulation Recognition And Its Dsp Realization

With the diversification and complexity of the modulation type of communication signals, the modulation classification is more and more important in shortwave wideband receiving system. This thesis studies the algorithm of digital modulation recognition for the commonly used digital signals such as OFDM,FSK,PSK and QAM, also that of the intra-class MQAM (M<=64), this paper then selects the well performed and easily realized algorithm in accordance with the objectives, and the modulation type classification is then realized on ADSP-TS201 chip. The research work is outlined as follows:1. The research of the modulation recognition for the multi-carrier digital signal OFDM and single-carrier digital signals(FSK,PSK and QAM). Utilize the Gaussian property of the OFDM signals, this paper introduces a algorithm based on a characteristic parameter Q (the difference between the mean which is the four times of the signal envelop and two times of the square of the mean which is the square of the signal envelop). This algorithm does not require any pre-verified information, its characteristic is easy to acquire and of great stability. Simulation results and the practical signal test prove that this method is effective. The study finally compares this proposed algorithm with the higher order cumulants and higher order moments algorithm, it is observed from the simulations that this algorithm yields better performance and lower computational complexity.2. The discussion on the intra-class recognition algorithms for the single carrier signals (FSK, PSK and QAM). This paper used the wavelet ridge to recognize the single intra signals,which has a good results. Then, this paper initially introduces the envelope variance recognition algorithm and another algorithm based on the he signal statistical parameters. Finally this paper compares the above mentioned 3 algorithms in white Gaussian channel and mid-range shortwave channels, it is observed that the wavelet ridge algorithm produces the highest recognition rate, however, it yields the most computational complexity.3. The analysis of the intra-class recognition algorithm for MQAM. The paper modified the existing algorithm with parameter Q to replace the sixth-order cumulants, this is due to the parameter Q has good recognition rate and requires a lot less computational effort. This paper also analyzes the performance of the method through simulation.4. Aiming at the problem on the frequency offset estimation of QAM signals; this paper used different ways to different channels. This paper initially takes the instantaneous frequency of the signal to roughly estimate the carrier frequency When the channel condition is good, the DFT phase different method can be utilized to carry out precise estimation, and this method produces high level of accuracy in decent and mid-range shortwave channels; however, when the channel condition is vile, we adopt the cyclic correlation method to solve the estimation problem on the carrier frequency because of the decent anti-noise ability and high level of accuracy. The thesis considers adequately the effect of shortwave channel and pulse shaping filter on algorithm, Simulation results a prove that this method is effective.5. The realization of the proposed modulation algorithm on the ADSP-TS201 chip. In accordance with the thesis requirement, a fully structured DSP module is established, and a series of measures have been taken to optimize the coding with concerns of coding and memory optimization. The choice of modulation type is realized on ADSP-TS201 chip fast and efficiently.