PrefaceThe engineering objective of high performance control using the tools of optimal control theory, robust control theory, and adaptive control theory is more achievable now than ever before, and the need has never been greater. Of course, when we use the term high performance control we are thinking of achieving this in the real world with all its complexity, uncertainty and variability. Since we do not expect to always achieve our desires, a more complete title for this book could be "Towards High Performance Control".To illustrate our task, consider as an example a disk drive tracking system for a portable computer. The better the controller performance in the presence of eccentricity uncertainties and external disturbances, such as vibrations when operated in a moving vehicle, the more tracks can be used on the disk and the more memory it has. Many systems today are control system limited and the quest is for high performance in the real world.In our other texts Anderson and Moore (1989), Anderson and Moore (1979), Elliott, Aggoun and Moore (1994), Helmke and Moore (1994) and Mareels and Polderman (1996), the emphasis has been on optimization techniques, optimal estimation and control, and adaptive control as separate tools. Of course, robustness issues are addressed in these separate approaches to system design, but the task of blending optimal control and adaptive control in such a way that the strengths of each is exploited to cover the weakness of the other seems to us the only way to achieve high performance control in uncertain and noisy environments.The concepts upon which we build were first tested by one of us, John Moore, on high order NASA flexible wing aircraft models with flutter mode uncertainties. This was at Boeing Commercial Airplane Company in the late 1970s, working with Dagfinn Gangsaas. The engineering intuition seemed to work surprisingly well and indeed 180 • phase margins at high gains was achieved, but there was a shortfall in supporting theory. The first global convergence results of the late 1970s for adaptive control schemes were based on least squares identification. These were harnessed to design adaptive loops and were used in conjunction with vi Preface linear quadratic optimal control with frequency shaping to achieve robustness to flutter phase uncertainty. However, the blending of those methodologies in itself lacked theoretical support at the time, and it was not clear how to proceed to systematic designs with guaranteed stability and performance properties.A study leave at Cambridge University working with Keith Glover allowed time for contemplation and reading the current literature. An interpretation of the Youla-Kučera result on the class of all stabilizing controllers by John Doyle gave a clue. Doyle had characterized the class of stabilizing controllers in terms of a stable filter appended to a standard linear quadratic Gaussian LQG controller design. But this was exactly where our adaptive filters were placed in the designs we developed at Boeing. Could we impro...

No abstract

This paper develops the adaptive disturbance estimate feedback schemes of a companion paper for enhancing the performance of controllers designed by off-line techniques. The developments are based on the parametrization theory for the class of all stabilizing controllers for a nominal plant, and the dual class of plants stabilized by a nominal controller. Such parametrization allows us conveniently to parametrize plant uncertainties for on-line identification and control purposes, minimizing the effects of unmodelled dynamics. Based on these parametrizations, along with prefilteringwhich minimizesthe effect of unmodelled dynamics, standard adaptive stabilization, adaptive pole assignment, or adaptive linear quadratic schemesare shown to achieve controller enhancement. The idea is to exploit a priori information about a plant and design objectivesin an off-line design, and yet exploit the power of adaptive techniques to learn and tune on-line. Attention is focused on techniques for fixed but uncertain plants. IntroductionControl theory has tended to develop in two separate directions. Off-line robust control exploits the a priori information about a plant and the off-line power of computers to achieve robust controllers that achieve performance objectives. On-line adaptive control theory has as its ideal to learn and implement on-line whatever is necessary to achieve the control objectives. Most of the significant results developed are those for the input-output (black box) model. Adaptive schemes are known to work well for low-order models with simple objectives. Inclusion of a priori plant information does not always allow a convenient plant parametrization for on-line identification based on least squares, although the less well understood recursive prediction error schemes can be applied.There is still a need for methods to apply adaptive techniques efficiently to assist in the control of high-order plants when there is a priori plant information. The challenge addressed here, as by Moore (1988) andWittenmark (1988), is to find convenient parametrizations which allow adaptive techniques to work.In our earlier paper (Tay and Moore 1988) the problem of enhancing a fixed controller performance using additional filtering and standard recursive least-squares based algorithms is developed. Off-line and on-line controller designs are blended harmoniously together based on the theory for the class of all stabilizing controllers. In this paper, the approach of Tay and Moore (1988) is broadened to permit the blending of standard adaptive pole-assignment, or adaptive linear quadratic designs, or indeed any adaptive stabilizing scheme to achieve enhancement of off-line designed

No abstract

No abstract