Modulation Classification In Multipath Fading Environments

Abstract

A new approach to classification of MQAM/MPSK signals in multipath fading environments is presented. The proposed approach, in which the two-step equalization strategy and higher-order cumulants based classifier are adopted, can effectively classify the MPSK and high-order QAM signals. The performance of proposed approach is evaluated by the computer simulations, which shows it has better classification ability. I. INTRODUCTIONModulation Classification of a received signal plays an important role in a variety of military and commercial applications. The examples include spectrum monitoring and management, surveillance and control of broadcasting activities, adaptive transmission schemes, and electronic warfare. Although significant progress has been made in the areas of modulation classification in flat fading environments, less attention is paid to modulation classification in multipath fading environments. In multipath fading environments, the inter-symbol interference (ISI) causes the distortion of received signal, which deteriorates the performance of classification methods developed for the flat fading channels.Among the existing methods, Lay[1] employed persurvivor processing (PSP) to address modulation classification with an assumption of a known multipath channel impulse. Flusser[2] used so-called fading invariants to classify the signals, but the construction of the fading invariants requires the channel impulse response must be symmetrical, which is impractical in the applications. Paris[3] presented a classifier that incorporated blind channel identification into universal classifier, and discriminated only between 4QAM and 8QAM. Wang[4] presented an approach that also uses the blind equalization to mitigate effects of multipath fading, then clusters signal spatial distribution, and the resulting centres are extracted to match the standard constellation patterns. Theoretically, the method can classify MASK, MPSK and MQAM signals, but the actual ability to classify is restricted by the performance of the blind equalization. This method, generally speaking, can only classify the signal with modulation order no more than 32. Xi[5] utilized higher-order statistics (HOS) to identify the channel, then the estimated channel was used to modify the higher-order cumulants based classifier for the flat fading channel. It should be noted that the classifier itself don't distinguish the signals with high order modulation. On the other hand, if the error of channel estimation is considered, the method performs more poorly.

Modulation Order Reduction Method for Improving the Performance of AMC Algorithm Based on Sixth–Order Cumulants

“…To the best of our knowledge, the only work that ever explored the idea of modulation order reduction for the purpose of AMC is the one described in [20]- [21], where authors proposed transformation of M-QAM signal y into the new M'-QAM signal y' by applying absolute value operation on the real part and the imaginary part of the original M-QAM signal:…”

“…The same can be concluded with any other AMC methods which are derived on the basis of statistical properties of AWGN: after transformation in equation 10, statistical properties of AWGN change dramatically and algorithms collapse. This would be the reason why the method proposed in [20]- [21] wasn't explored by other authors in significant extent. In order to preserve statistical properties of both QAM signal and AWGN, along with achieving modulation order reduction, we propose the following method:…”

“…which provides continuous errorless classification even slightly below SNR=10dB. It should be noted that this does not stand in the case of using method [20]- [21]. Only preserving both statistical properties of transmitted signal and noise simultaneously, leads to this kind of result.…”

KNN Based Classification of Digital Modulated Signals

“…Modulation classification with wavelet coefficients has better resistance to the noise at the cost of higher complexity, compared with that with instantaneous information. HOS is a more frequently adopted feature than the previous two, generally including the statistics of moments and cumulants [16][17][18]. In addition to classification of the modulation types, HOS can distinguish modulation orders with high resistance to the additive white Gaussian noise (AWGN) as well.…”

Optimal Decision Fusion Based Automatic Modulation Classification by Using Wireless Sensor Networks in Multipath Fading Channel

“…New features based on sixth-order cumulants were proposed for classification of MPSK signals in [21]. A high-order cumulants based AMC approach has been proposed in [22] to overcome the channel mismatch impacts of multipath fading channels with a two-step equalization strategy.…”