Personally, I am very interested in the mapping between digital filter realizations and optimal control problems. I am amazed by what a low-order recursive filter can do with just a handful of delay, multiply, and add, (DMA) operations. Many real feedback control systems also have very nice low-order representations in continuous time. One approach is to discretize a continuous-time solution; another approach is to formulate the problem/solution directly, as a digital filter realization. I think the first approach is generally the most popular approach, so a lot has been achieved in that area. I think less has been done in the latter, and I believe that it is also the most useful. The trick here is to know how each digital DMA component (implemented on/in a computer) affects the outside world via the feedback loop (in which the computer is placed).              

Thanks Hugh,

Can you please elaborate a little more on the second approach? 

Biswajit

I wrote up my latest ideas on this in a conference paper

https://www.researchgate.net/publication/299569076_Optimal_Design_of_Digital_Anticipatory_Servomechanisms_with_Good_Noise_Immunity

that I will be presenting next month at

http://21stcenturywiener.org/

I hope it makes sense; otherwise, I'd be happy to clarify anything. 

Conference Paper Optimal Design of Digital Anticipatory Servomechanisms with ...