In previous work, we have discussed the use of operators to solve some signal analysis problems that are at the heart of some signal processing and analysis problems. In this paper, we extend the formalism to include noise by the introduction of a noise operator as an alternative way to think about the various signal in noise problems in radar and sonar. In addition, we show how non-linear combinations of signal with noise can be analyzed as noise operators acting on signals can expressed as combinations of operators with random parameters with a given distribution acting on the signal. Then, we show that the noise reverberation or clutter problem can be thought of as a parameter randomized operator acting on signals in either the radar and sonar domain. Then, we note that this formulation lends itself to finding unexpected consequences of non-linear transformations of random variables. Finally, we note a radar problem first addressed by Cartwright and Littlewood in WWII of identifying a spectrum being produced by a non-linearity in a circuit versus noise can be attacked by these methods.
|