Decision directed estimation of the state sequence of an unknown Markov Chain

We address the simultaneous estimation of the state sequence of a discrete-time finite-state Markov chain and the unknown probabilities governing the Markov chain. The availability of discrete observations made in the presence of memoryless noise is assumed. We describe a decision-directed estimator based on the Viterbi algorithm. Simulation results are presented which indicate that the decision-directed estimator´s performance compares favorably to the Viterbi algorithm using the true Markov chain probabilities.