Automatic Modulation Recognition of Digital Communication Signals

The modulation recognition technology of communication signals has been an important theme of wireless communication. Based on the parameters abstraction of time domain statistical feature and fractal feature, the feature vector samples is formed. The artificial neural network is the research hot spots of pattern recognition. An artificial neural network is proposed for an automatic recognition of different types of digital pass-band modulation. The feed-forward networks are trained to recognize 2ASK, 4ASK, 2FSK, 4FSK, BPSK, QPSK, l6QAM, 64QAM signals with better generalization as well as an addition of a new statistical features set. Performance of the processor in the presence of additive white Gaussian noise (AWGN) is simulated. The experiments show that comparing with traditional methods the network model and training algorithm designed in this paper is improved much in convergence speed, training time and recognition ratio. Simulations show satisfactory results even with low value, e.g. 98% success rate at 8dB SNR.