Statistical analysis of power spectra of signals governed by Markov chains

The paper presents a novel derivation of the average autocorrelation and the average power spectrum for Markov renewal processes. Motivated by a practical problem arising in randomized switching in power electronic circuits, we provide a linear filtering interpretation of the process that forms continuous switching waveforms by concatenation of segments that are associated with various states of a Markov chain. This new, streamlined derivation of formulas for autocorrelation and power spectrum is of interest for optimized synthesis of randomized switching waveforms.