# Method to find Correlation for a random variable and thereby power spectral density of random signal

I'm trying really hard to understand the segment of text from a standard communication book ( Modern digital and analog communications by B P Lathi). We are trying to find the power spectral density of binary signalling where the value can take random values 1,0,-1 with probabilities given. What I understand is that to find the power spectral density the very first step is to find Correlation of the random values . I'm not able to understand the method . I understood how the mean and variance is found , but not how Correlation for 1 is found and why does Correlation for greater than or 2 be zero. Could someone please explain an alternate simple method or bother to please explain the same method ?

• which book is that? Some of us might have it, and knowing we can just look inside might make things easier to explain, because we could use the book's whole way of thinking things. – Marcus Müller Aug 21 '18 at 19:49
• FWIW bipolar encoding is not binary encoding but is defined here en.m.wikipedia.org/wiki/Bipolar_encoding – Tony Stewart Sunnyskyguy EE75 Aug 22 '18 at 2:14
• @Marcus Muller , the book is book ( Modern digital and analog communications by B P Lathi – Fawaz Aug 22 '18 at 3:54