时间平均和系综平均.pdf

Chapter 1 Mathematical Methods In this chapter we will study basic mathematical methods for characterizing noise pro- cesses. The two important analytical methods, stochastic functions and Fourier analysis, are introduced here. These two methods will be used frequently throughout this text. 1.1 Time Average vs. Ensemble Average Noise is a stochastic process consisting of a randomly varying function of time and thus is only statistically characterized. One cannot argue a single event at a certain time; one can only discuss the averaged quantity of a single system over a certain time interval or the averaged quantity of many identical systems at a certain time instance. The former is called time average and the latter ensemble average. Let us consider identical systems which produce noisy waveforms () (), as shown in Fig. 1.1. Figure 1.1: Ensemble average vs. time average. 1 One can define the following time-averaged quantities for the -th member of the en- semble: () () = lim 1 2 () () 2 (mean = first-order time average) (1.1) () 2 1 2 () 2 () = lim () 2 (mean square = second-order time average) (1.2) () ( ) () ()() ( + ) = lim 1 2 () ()() ( + ) 2 (autocorrelation function) (1.3) One can also define the following ensemble-averaged quantities for all members of