Non-cooperative Communication Demodulation Key Technologies

Non-cooperative communications refers to an unauthorized third party accessing the cooperative communication systems without interfering the normal message delivery for cooperative communications. Parameter estimation and demodulation for non-cooperative communications mean that the non-authorized third party uses modern digital signal processing technology to estimate the parameter of the received signal and then complete demodulation.Based on non-cooperative communications, this dissertation makes a deep research on its key parameter estimation and demodulation algorithms, including the carrier recovery algorithm, the symbol rate estimation algorithm and the modulation recognition algorithm. The dissertation is divided into five parts, the above three kinds of algorithms are introduced in the second, third and fourth part respectively.The first part is the introduction, which introduces the significance of the parameter estimation and demodulation algorithms for non-cooperative communication systems firstly, and then describes the basic structure of non-cooperative communication systems and the development of related algorithms, finally introduces the main contents, structural arrangement, contributions and innovation of this dissertation.The second part mainly studies the carrier recovery algorithm. This part proposes a new structure of carrier recovery loop for the QAM receiver. In this structure, the frequency detector and the phase detector are paralleled; meanwhile a modified weighted Decision-Directed(MWDD)algorithm is adopted in phase detection. It has been verified by simulation that the algorithm improves acquisition rate of carrier recovery loop and reduces the phase jitter. The 16-QAM system can acquire carrier quickly even up-to 10% frequency offset of the symbol rate.The third part mainly investigates the symbol rate estimation algorithm. Based on the localization ability in time-domain and frequency-domain, wavelet transform has the excellently transient detection ability when a symbols changes. The existing Haar wavelet symbol rate estimation algorithms assume that the symbol is a rectangular pulse, without considering the influence of pulse shaping filter. In order to solve the above problems, this chapter takes into account the pulse shaping filter technology impacting on the signals, uses Daubechies (dbN) wavelet for band-limited communication system to extract transient characteristics of symbol change and then estimates MPSK symbol rate with FFT algorithm. Simulation results indicate that the proposed symbol rate estimation algorithm is better than classical Haar wavelet algorithm.The fourth part mainly researches the modulation recognition algorithms. Neural network is an adaptive nonlinear signal processing systems, which has a strong capability to identify the pattern classification. This part identifies MPSK modulation type based on radial basis function (RBF) neural network and higher order cumulating. A new variable step size factor for LMS algorithm is proposed, which makes the actual output more close to the expected one. Simulation results show that the method improves the identification rate effectively.The Fifth part is a summary of the whole dissertation and pointed out the further research work in the figure.