Simplified algorithms for self tuning control

The simplification of self-tuning controllers have been considered here. Firstly, approximations to full parameter estimation schemes can be used results from this showed that in practice the simplified schemes work just as well as their complicated counterparts, so it seems reasonable to use the simplified procedure for the majority of systems. Model reduction methods are only really useful on-line when (a) the system is of high order (4th order or above) and/or (b) the system exhibits both high and low frequency dynamics, use of a standard low order self-tuning controller in these cases could result in, at best, a poor control action being applied. Finally for many systems a detuned but fairly robust control objective is much more desirable than an optimal control scheme which can be quite temperamental. Both PID and deadbeat controllers fall into this category and it has been shown how self-tuning controllers can be derived, in many different ways, in order to achieve such goals.