A proposed digital modulated signal identification based on pattern recognition

Digital modulation identification algorithm based on statistical pattern recognition techniques, has been developed to automatically identify the modulation type of a fixed length signal segments, and intercepted by fixed bandwidth HF radio receiver. The proposed algorithm is developed by generating and statistically analyzing a large set of numerically simulated signals. The resultant algorithm is successfully tested on another independent set of simulated signals. The developed algorithm allows automatic identification for the following digital modulated signal types: FSK-2, PSK-2, FSK-4, and PSK-4. The algorithm is developed to be robust under wide variations of operating conditions, including signal to noise ratio, message content, baud rate, tuning error, data acquisition system quantization noise, and signal interference. Algorithm performance is tested and investigated with the presence of some expected perturbing factors such as noise, and mismatch of the center frequency.