TABLE OF CONTENTS In this chapter we learn how to generate random number sequences, noise, and clutter. All signals are invariably contaminated by noise and/or clutter. The contaminated signal is filtered or estimated to extract the signal by digital signal processing. Filtering will be covered in Chapter 3, the time-domain filtering by finite impulse response (FIR) or the infinite impulse response (IIR) filter. Filtering in the frequency domain will be covered in Chapter 4: fast Fourier transform (FFT) and inverse fast Fourier transform (IFFT). Estimation of signals will be covered in Chapter 8: Kalman filter.
There is no better way to understand signal processing required than by knowing the characteristics of noise or clutter and no better way to understand noise and clutter than by generating them ourselves so that we have a thorough understanding of the noise and clutter we will encounter in the real world.
The kind of noise and clutter we have heard of, such as Gaussian noise, Ray-leigh noise, exponential noise, chi-squared noise, lognormal clutter, and Weibull clutter, can be generated from unit uniform random variables. Unit uniform random variables are, in turn, generated from random number sequences.
True noise, a stochastic process, can never be manufactured (or generated) by a deterministic machine such as computer. In other words, true random number sequences can never be generated by computer programs, since the true random numbers do not repeat, do not have cyclic periods, do not have an end. However, we can generate a random number sequence as close...
