Communication waveform performance prediction in partial-band partial-time jamming RF environments

In complex radio frequency (RF) environments with partial-band and partial-time (PBPT) jamming, the Signal-to-Noise Ratio (SNR) of a communication link may have multi-modal distributions over a range of SNR levels. The control of transmission powers in a communication system is able to shift the link SNR distribution up or down without altering the distribution's shape. However, it is important to determine how much a link SNR distribution should be shifted such that a communication waveform is able to operate. In this paper, a neural network (NN) based method is proposed to evaluate the required SNR shift for a waveform to operate on a link with complex multi-modal SNR distribution. The NN inputs are derived from a normalized SNR distribution, which is a shifted version of the “actual” (measured) link SNR distribution. The NN output is the required minimum SNR shift of the normalized SNR distribution, i.e., a normalized SNR shift, for the waveform to yield Bit-Error-Rates (BERs) less than a specified maximum acceptable rate. The NN may be trained with training data samples obtained from simulations or emulations. After training, the NN is shown to be able to accurately evaluate the required normalized SNR shift for a range of normalized link SNR distributions corresponding to various complex communication link SNR conditions. Based on the NN's evaluations, accurate transmission power control and/or waveform selection can be achieved in communication systems operating in complex PBPT jamming RF environments, which lead to superior communication performance and robustness.