Analysis of nonlinear stochastic systems with wide-band inputs

A general method for approximating systems with wide-band inputs by diffusion processes is discussed. The input noise itself can be processed nonlinearly by elements of the system. One particular case is examined in detail, that of a hard limiter just after the input. The results suggest that the limiter can actually improve performance if the noise intensity is small. The development illustrates how tricky, but potentially useful nonlinearities can be handled when the system input is wide-band noise.