Turbo codes for time varying AWGN channels with application to FH/SSMA

We investigate the application of parallel concatenated convolutional codes (i.e., turbo codes) to additive white Gaussian noise channels with time-varying variance. Applications of this model include multiple-access interference and partial-band noise jamming in frequency hopped networks. Three schemes for evaluating the likelihood ratio of the channel output are compared. These schemes provide tradeoffs between simplicity of implementation, BER performance, and robustness to inaccuracies in the measurements.