Abstract
The development of Fourth Generation (4G) wireless communication systems is anticipated to heavily rely on software radio technology. One of the main responsibilities of an intelligent receiver is the signal identification process, which is a stage in between signal interception and demodulation. The method of determining what modulation is being used at the transmitter based on observations of the received signal is known as Automatic Modulation Classification (AMC). Ten candidate signals were produced, including 2ASK, 4ASK, 2PSK, 4PSK, 2FSK, 4FSK, 16 QAM, GMSK, 64QAM, and 256 QAM. AWGN and the multipath Rayleigh fading effect were simulated in order to mimic channel conditions. Initially, instantaneous characteristics including amplitude, phase, and frequency were obtained. Instantaneous features were used to create stochastic features. The classifier was developed using seven essential characteristics. Higher order statistical features like moments and cumulants were used to classify higher order QAM signals like 64QAM and 256QAM. Threshold values were used in the development of the Decision Tree classifier. For SNR=3dB, the overall classification result was more than 97%. For SNR=5dB, the success rate was almost 99% (no fading condition). Ten modulated signals with different channel characteristics and SNRs as low as -5dB may be classified using the created classifier.