Performance studies of slow frequency hopping M-ary FSK with concatenated codes in partial band noise jamming

The performance of three channel coding alternatives with slow frequency hopping non-coherent M-ary FSK is compared. The channel is considered as an AWGN-channel with partial band noise jamming. The coding alternatives chosen for comparison are convolutional (7, 1/3) -code, concatenation of convolutional (7, 1/3) and Reed-Solomon (63, 31)-code and concatenation of convolutional (7, 1/3)- and Golay (23, 12)-code. Hard decision metric without jammer state information (JSI) is used. For wideband jamming the results show that for moderate bit error probabilities, concatenation of codes does not provide coding gain. For moderate values of bit error probabilities the convolutional code gives the best performance. For low values of bit error probabilities the combination of convolutional and Reed-Solomon coding is the best alternative. Results obtained by Monte Carlo-simulations and quasi-analytical calculations show that channel coding efficiently mitigates the effects of partial band jamming. For small fractions (<0.2) of jammed FH bandwidth, concatenation of codes removes error efficiently and thus forces the jammer to use larger jamming fractions.