Time synthesis of servo systems

Until recently, controller synthesis could not be performed directly in the time domain. We had recourse to so-called best-practice methods (the most famous of which is perhaps the Ziegler-Nichols method) which, from readings of characteristic system parameters (such as those provided by its index response, for example), produced parametric values of classical controllers of the P, P + I or PID type. With these parameters, we could run different simulations, to test their relevance to different criteria. We could then validate them, or repeat the whole procedure to try and find better ones.

Since the advent of personal computers, and the high-performance numerical algorithms that accompany them, we've had the opportunity to reverse the...