FHSS Signals Classification by Linear Discriminant in a Multi-signal Environment

Frequency-hopping spread spectrum (FHSS) spreads the signal over a large bandwidth where the carrier frequencies change quickly according to a pseudorandom number making signal classification difficult. Furthermore, classification becomes more complex with the presence of additive white Gaussian noise (AWGN) and interference due to background signals. In this paper, a linear discriminant (LD) method based on the Euclidean distance is proposed for the classification of FHSS signals in the presence of AWGN and background signal. Probability of correct classification (PCC) of the FHSS signals is performed by the LD method for the signal-to-noise ratio (SNR) range of −6 to 15 dB. Results show that the proposed method has achieved 90% detection rate at the SNR range of −1.6 to 3.5 dB in the presence of AWGN only, while its performance is degraded to 0.9 to 12 dB when the background signal is present.