Linear quadratic Gaussian control has been studied extensively for relatively low order stochastic processes. However, for highly random, unobservable processes. The LQGC gives a poor performance, especially for nonlinear MIMO systems. So, could developing controllers for such a process be built?
I am not shure what you mean by "highly random unobservable processes". Off course, if you want to control a system, you have to observe it, i.e. find a measureable quantity In [1] you find an example where we controlled the penetration depth of laser welding processes - a highly complex and nonlinear process - by the frequency of the appearance of an image feature (the 'full penetration hole'). What you have to clarify is the (physical) relation between the the measured variable and the manipulated variable. This can be done by modelling and/or by open-loop measurements. If you have highly nonlinear systems, you have to care about hysteresis effects and time-dependencs.
I hope I could help you.
[1] Blug, A.; Abt, F.; Nicolosi, L.; Heider, A.; Weber, R.; Carl, D. et al. (2012): The full penetration hole as a stochastic process: controlling penetration depth in keyhole laser-welding processes. In: Appl. Phys. B 108 (1), S. 97–107.
Thanks, Blug,
I read the references however, I think you're right. I've been trying different methods. But, to give more clear descriptive talk, the problem here considering not only the control action but also the coupling between different states. and all this states are noisy. So, for this type of process what's the confined procedures for designing the controller.
- Badges
- Science topic