1: \begin{abstract}

2: %In this paper the convergence properties of errors in adaptive systems are explored. Building upon the work in \cite{jen13} the analysis is expanded to a higher order system with limited persistence of excitation. Similar to the first order systems, an infinite region in the state-space of the adaptive system where the state rate is bounded, and $e$ is bounded away from the origin for a finite time $T_1$ which is identified. The region of slow convergence of the adaptive system in \cite{jen13} is linked to an attenuation of the PE of the input $r$ to $x_p$ which is the signal that drives the system. This attenuation is a function of system's internal parameters and is the primary reason exponential stability cannot be achieved, at best uniform asymptotic stability in the large can be shown.

3: %

4: %%In this paper we explore the convergence properties of errors in adaptive systems. We show that there exists an infinite region in the state-space of adaptive systems where the state rate is bounded. This region of slow convergence can be tied to the level of persistence of excitation of the plant state which is not only a function of the reference input $r(t)$, but, the parameter error $\widetilde \theta(t)$ as well. %In adaptive systems with more than one adaptive parameter the rate of convergence will very much depend on reference input, yielding fast convergence of the state error and no convergence for the parameter errors without sufficient persistence of excitation, or, slow convergence in all errors if persistent excitation conditions are sufficient.

5: %%This dependency of persistence of excitation on the system's internal parameter error is a primary reason exponential stability cannot be achieved, at best uniform asymptotic stability in the large can be shown.

6: %\end{abstract}

7: