On the other hand, the discrete tim

On the other hand, the discrete time algorithms are usually viewed as a combination of a parameter estimator (driven by a suitable prediction error) and a certainty equivalence control law. The former approach is known to have certain difficulties including the fact that, in general, the error y - y* must be augmented by other terms to guarantee stability, and strict positive real (SPR) conditions on various transfer functions appear as necessary conditions for convergence. Also, it is difficult to see what properties are retained if the control law is changed, e.g. if the input reaches a saturation limit. On the other hand, most of the discrete algorithms retain the key properties of the parameter estimates irrespective of the control signal.