Optimal Control over Multiple Input Lossy Channels

TL;DR

This paper analyzes how different communication protocols (UDP-like vs TCP-like) affect the optimal control performance over multiple lossy input channels, showing UDP-like protocols lead to higher costs.

Contribution

It extends previous work to multidimensional, multi-actuator systems, providing a proof that UDP-like protocols result in larger quadratic costs in optimal control.

Findings

UDP-like protocols increase control costs compared to TCP-like protocols.

The cost difference between protocols can be quantified and analyzed.

Numerical examples demonstrate the impact of protocol choice on control performance.

Abstract

The performance of control systems with input packet losses on the controller to plant communication channel is analysed. The main contribution of this work is a proof that linear optimal control systems operating with UDP-like communication protocols have a larger quadratic cost than the same systems operating with TCP-like protocols. The proof is derived for the general case of multidimensional and independent actuation communication channels. In doing so, our results extend previous work to systems with multiple distributed actuators. The difference in cost between two communication protocols is analysed, enabling the maximal difference between the two protocols to be quantified. Numerical examples are presented to highlight the difference in costs induced by the choice of communication protocol.