A Norm Optimal Iterative Learning Control framework towards Internet-Distributed Hardware-In-The-Loop simulation

Internet-Distributed Hardware-In-the-Loop (ID-HIL) simulation couples geographically-dispersed hardware-in-the-loop setups into an integrated system simulation over the Internet. The challenge is that the Internet is a dynamic system on its own with delay, jitter, and loss that can interfere with the dynamics of the integrated system, lowering the simulation fidelity. This paper focuses on the effects of delay in a linear time invariant ID-HIL system and presents a Norm Optimal Iterative Learning Control (NO-ILC) based method to improve the fidelity of the integrated system. It is shown that the method guarantees monotonic convergence; i.e., with each iteration, the fidelity of the simulation will improve further. A recursive, data-driven calculation method is also presented that avoids the need for ad-hoc tuning of learning gains and does not require any knowledge about the setups that are being integrated. A simulated networked mass-spring-damper system is used to demonstrate the technique.