Part VII: MIMO Control Design
We next examine techniques specifically aimed at MIMO control problems in which interaction cannot be ignored if one wants to achieve maximal performance. The first chapter in this part covers linear optimal control methods. These techniques are very frequently used in advanced control applications. Indeed, we will describe several real world applications of these ideas. The next chapter covers Model Predictive Control. This is an extension of optimal control methods to incorporate actuator and state constraints. These ideas have had a major impact on industrial control especially in the petrochemical industries.
The final chapter covers fundamental design limitations in MIMO control. As in the SISO case, MIMO control system design involves an intricate web of trade-offs. Indeed, many of the core issues that we have seen in the case of SISO systems have direct MIMO counterparts, e.g., poles, zeros, sensitivity functions, disturbances, robustness, etc. However, unlike SISO systems, these issues now have a distinctly directional flavor, i.e., it turns out that it matters which combination of inputs and outputs are involved in a particular property. These directional issues require us to be a little more careful in analysis, synthesis and design of MIMO control loops. This thus sets the scene for the next five chapters.