Improvability of feedback systems (Q1593781)

This paper introduces the notion of improvability and bottlenecks of feedback systems: A feedback system is improvable if its performance can be enhanced by re-allocating sensor and actuator costs under budget constraints. Linear SISO feedback systems with quadratic performance index and LQG controller equations are considered. The following problems are given: 1. Derive a rule for determining whether a given system is improvable under constraints or not. 2. Derive a rule for identifying the bottleneck of a feedback system. A result proved in this paper is that both improvability and bottleneck can be evaluated using on-line measurements in the feedback loop, the knowledge of the plant data is not necessary. Four examples illustrate the results. The authors hope to generalize the ideas to more complicated models.