A real-time design of a detection and classification system for narrowband and broadband systems

The authors present a real-time signal processing system which incorporates a pattern recognition expert system for detecting narrowband and broadband signals in the presence of noise. The system-based on the AT&T Bell Laboratories DSP32 Microprocessor-detects and classifies signals by extracting parameters of the power spectrum. From the extracted shape description, decisions are made on the basis of the presence/no presence of narrowband and broadband signals. The system can detect signals from either use-defined parameters or a time-sequence sample of the desired sequence. The system was tested on real data at 0-4 kHz. Using unshaped and shaped white noise, the system successfully detected narrowband signals down to -28-dB SNR (broadband noise dominant) and -32-dB SNR (narrowband noise dominant), respectively