非协作通信解调关键技术研究

【作者】 李小娟；

【导师】 朱立东；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2010， 硕士

【摘要】 非协作通信是指未经授权的第三方,在不影响协作通信双方正常通信的前提下接入到协作通信系统中的通信方式。而非协作通信的参数估计和解调是指非授权的第三方利用现代数字信号处理技术完成对接收信号的调制参数估计以及完成后继的解调工作。本文就是在非协作通信的背景下,对于其关键的参数估计算法和解调算法作了较为深入的研究,其中包括符号率估计算法,调制方式识别算法和载波恢复算法。全文共分为五个部分,各种算法的研究主要分布在第二、三、四部分。第一部分为绪论,在绪论部分主要介绍了对于非协作通信参数估计和解调算法研究的意义、非协作通信系统的基本结构和特点以及相关算法的发展,最后介绍了本文研究的主要内容、结构安排和贡献及创新。第二部分主要研究了载波恢复算法。在本章中提出了一种新的QAM载波恢复环路结构。在此结构中频率检测和相位检测同时进行,在鉴相阶段引入了改进的加权直接判决(MWDD)算法。经仿真实验验证此方法提高了整个载波恢复环路的捕获速度,并使载波环路的相位抖动有所减小,16-QAM载波环路可以捕获10%符号率的频偏。第三部分主要研究了符号率估计算法。小波在时域和频域具有优良的局部化特征,适用于信号瞬时变化特征提取。现有的使用Haar小波的符号率估计算法中都假设码元是矩形脉冲,没有考虑实际通信系统脉冲成形滤波技术的影响。针对上述问题,本章在带限系统的模型下,考虑了脉冲成形滤波技术对信号产生的影响,使用适合带限系统信号瞬时变化特征提取的具有高阶消失矩的Daubechies(dbN)小波,结合FFT算法,提出了一种MPSK符号率估计方法。仿真结果表明,使用该符号率估计方法要比使用Haar小波性能更优越。第四部分主要研究了调制方式识别算法。神经网络作为一种自适应的非线性信号处理系统,具有强大的模式分类识别能力。本章使用基于径向基函数(RBF)神经网络分类器算法和高阶累积量进行MPSK调制类型识别,在RBF的输出层训练中,提出一种新的变步长因子的LMS算法,使实际输出更好地逼近期望输出。仿真实验表明,该方法有效地提高了模式正确识别率。第五部分为对全文的总结,并指出需要进一步研究的工作。更多还原

【Abstract】 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.更多还原