Information-theoretic bounds on trellis complexity of nonlinear codes

Information-theoretic measures are used to derive both upper and lower bounds on the state complexity and on the branch complexity of nonlinear block codes To the best of our knowledge, the proposed upper bounds are the only upper bounds that address nonlinear codes. The improvement of the novel lower bounds is twofold: these bounds are tighter than the existing bounds and they also bound the trellis complexity of improper and unobservable trellis diagrams, unlike the existing bounds.