Computer recognition of v cv speech units based on linear and non linear dynamical system models using brain like computing and statistical learning algorithms