A Viterbi detector for the acquisition of convolutionally encoded non-coherent MFSK

Addresses the fast synchronisation of frequency hopping waveforms in conditions where there is significant time-varying frequency error compared to the decision time of the detector. To ensure good error rate performance in the presence of burst interference, use is made of time-frequency diversity and convolutional forward error correction. The possibility of performing a relatively long-term integration (several transmitted data symbols) of the transmitted signal was investigated. A serial search is used and use is made of the redundancy in the error correction coding to perform a maximum likelihood (ML) detection of the presence of a signal. The results of analysis of the detection and false alarm probabilities in the presence of Gaussian interference is presented along with simulated results