Hop-timing estimation for FH signals using a coarsely channelized receiver

This paper addresses new hop-timing (epoch) estimation schemes which employ a coarsely channelized preprocessor in order to suppress the frequency and phase dependence in random frequency-hopping (FH) signals. Coarse channelization implies a bank of filters, covering the total spread bandwidth of the FH signal, whose number is much smaller than the size of the candidate hop frequency set. Linear and nonlinear combinations of preprocessor outputs are explored and compared. It is found that post-processing, rather than the size of the filter bank alone, is the determining factor on estimation performance. Performance evaluation is presented via both extensive computer simulations and analytical lower bounds. Comparisons with existing optimal and suboptimal systems are also provided.