Possible new directions in control

One has a distinct feeling that control theory is approaching a stalemate in the sense that one no longer sees contributions in the literature, which represent large strides. Rather, one has the feeling that the field is jogging in place. This state of affairs can be attributed to two factors. (i) There has been a great deal of very capable work over the last decade, that has thoroughly investigated the "canonical" questions raised in the fifties and early sixties. To the extent that some of these questions could be answered, they were answered, and those that remain open are largely intractable. (ii) We appear to be locked into a set of "canonical" problems and there has been little, if any attempt to raise new ones. A symptomatic area in this respect is pole placement. There has been a great deal of effort devoted to the canonical question of minimal compensators needed for arbitrary pole placement, and hardly any progress has been attempted on the unspoken next question which is that of how these pole placement results should be utilized in feed-back control system synthesis. In fact, it seems certain that the interesting results of the future should deal with synthesis. It is an area of great and largely unexplored possibilities. Among these: (i) techniques for interactive, computer aided: design; (ii) design techniques based on the use of trade off characteristics; (iii) the use of minicomputers in variable structure sampled data feedback systems; (iv) extensive use of optimization algorithms in the design of nonlinear feedback systems, partly as a tool for coping with nonlinearities and partly (simultaneously) with constraints in the time and frequency domains; (v) the specification of new types of control systems based on different performance expectations from those in current use. In short, it only seems that we are in a stalemate. By moving on to a new wave of exciting and meaningful research activity in the control area.