Markov processes asymptotically achieve the capacity of finite-state intersymbol interference channels

Magnetic recording channels are generally modeled as finite-state linear intersymbol interference (ISI) channels with additive white Gaussian noise and a binary input constraint. Lower bounds on the channel capacity have been computed by using this technique to estimate the information rates of optimized, order-r Markov input processes. RLL-constrained binary symmetric channel (BSC) and for two-dimensional finite-state ISI channels is proved in this paper