本文關(guān)鍵詞： 調(diào)制識(shí)別 特征提取 小波變換 高階累積量　出處：《長(zhǎng)春理工大學(xué)》2016年碩士論文　論文類型：學(xué)位論文

【摘要】：在現(xiàn)代通信技術(shù)中,通信信號(hào)的制式識(shí)別在電子對(duì)抗、頻譜管理、自適應(yīng)接收、認(rèn)知無(wú)線電等多個(gè)領(lǐng)域有廣泛的應(yīng)用。在現(xiàn)有的文獻(xiàn)中,通信信號(hào)的調(diào)制制式識(shí)別方法主要分為兩種:基于決策理論的最大似然假設(shè)檢驗(yàn)方法和基于特征提取的模式識(shí)別方法。第一種方法能夠達(dá)到很好的識(shí)別效果,但是計(jì)算復(fù)雜,而且對(duì)頻偏、相偏、定時(shí)誤差等非常敏感。相比較而言,第二種方法計(jì)算復(fù)雜度低,效率高,在特征選取合理的時(shí)候能夠達(dá)到很好的識(shí)別效果,且具有較高的穩(wěn)定性。因此本文重點(diǎn)研究基于特征提取的模式識(shí)別方法。本文應(yīng)用小波變換方法,對(duì)PSK、ASK、FSK信號(hào)提取了特征參量并對(duì)信號(hào)進(jìn)行了有效的分類。在假定噪聲為加性高斯白噪聲的條件下建立了信號(hào)模型,說(shuō)明了尺度因子的選取問(wèn)題,通過(guò)計(jì)算小波變換系數(shù)幅值提取了信號(hào)的特征參數(shù)。通過(guò)MATLAB對(duì)通信信號(hào)進(jìn)行了仿真,說(shuō)明了算法的有效性。本文應(yīng)用高階累積量,針對(duì)多徑信道條件,對(duì)BPSK信號(hào)和QPSK信號(hào)的制式識(shí)別進(jìn)行了較深入研究。在無(wú)線通信中,實(shí)際信道通常有多徑衰落現(xiàn)象。在這種情況下,建立在高斯白噪聲信道模型上的調(diào)制識(shí)別算法的性能通常會(huì)下降甚至失效。針對(duì)此問(wèn)題,提出了基于四階累積量和六階累積量相結(jié)合的調(diào)制識(shí)別算法。多徑數(shù)目為2時(shí),從理論上證明了該算法和基于四階累積量的算法相比,能夠更好的抗多徑干擾。仿真結(jié)果表明,在多徑衰落條件下,該算法對(duì)BPSK信號(hào)和QPSK信號(hào)的識(shí)別率高于基于四階累積量的算法。在信噪比為2dB的多徑衰落信道情況下,分類BPSK和QPSK信號(hào)的識(shí)別率幾乎達(dá)到100%;在識(shí)別BPSK信號(hào)時(shí),此算法性能明顯優(yōu)于基于四階累積量的算法。
[Abstract]:In modern communication technology, communication signal recognition has been widely used in many fields, such as electronic countermeasure, spectrum management, adaptive reception, cognitive radio and so on. The modulation recognition method of communication signal is divided into two kinds: the maximum likelihood hypothesis test method based on decision theory and the pattern recognition method based on feature extraction. The first method can achieve good recognition effect. But the calculation is complex and sensitive to frequency offset, phase offset and timing error. Compared with other methods, the second method has low computational complexity and high efficiency, and can achieve a good recognition effect when the feature selection is reasonable. Therefore, this paper focuses on the research of pattern recognition based on feature extraction. In this paper, we apply wavelet transform method to PSK ask. The characteristic parameters of FSK signal are extracted and the signal is classified effectively. The signal model is established under the assumption that the noise is additive Gao Si white noise, which explains the selection of scale factor. The characteristic parameters of the signal are extracted by calculating the amplitude of the wavelet transform coefficient. The simulation of the communication signal by MATLAB shows the validity of the algorithm. In this paper, the high-order cumulant is applied. In view of the multipath channel condition, the standard recognition of BPSK signal and QPSK signal is deeply studied. In wireless communication, the actual channel usually has multipath fading phenomenon. In this case. The performance of modulation recognition algorithm based on Gao Si white noise channel model usually decreases or even fails. A modulation recognition algorithm based on the combination of fourth-order cumulant and sixth-order cumulant is proposed. The number of multipath is 2:00. It is proved theoretically that this algorithm is compared with the algorithm based on fourth-order cumulant. The simulation results show that under the condition of multipath fading, it can resist multipath interference better. The recognition rate of BPSK signal and QPSK signal is higher than that based on fourth-order cumulant. In the case of a 2dB signal-to-noise ratio (SNR) multipath fading channel. The recognition rate of classified BPSK and QPSK signals is almost 100. The performance of this algorithm is better than that based on fourth order cumulant in the recognition of BPSK signals.
【學(xué)位授予單位】：長(zhǎng)春理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TN911.3
，

本文編號(hào)：1444208