Robust Control Techniques for State Tracking in the Presence of Variable Time Delays.

摘要

In this paper, a distributed driver-in-the-Ioop and hardware-in-the-Ioop simulator is described with a driver on a motion simulator at the U.S. Army TARDEC Ground Vehicle Simulation Laboratory (GVSL). Realistic power system response is achieved by linking the driver in the GVSL with a full-sized hybrid electric power system located 2,450 miles away at the TARDEC Power and Energy Systems Integration Laboratory (P&E SIL), which is developed and maintained by Science Applications International Corporation (SAIC). The goal is to close the loop between the GVSL and P&E SIL over the Internet to provide a realistic driving experience in addition to realistic power system results. In order to preserve a valid and safe hardware-in-the-Ioop experiment, the states of the GVSL must track the states of the P&E SIL. In a distributed control system utilizing the open Internet, the communications channel is a primary source of uncertainty and delay that can degrade the overall system performance and stability. The presence of a cross-country network delay and the unavoidable differences between the P&E SIL hardware and GVSL model will cause the GVSL states and P&E SIL states to diverge without any additional action. Thus, two robust strategies for state convergence are developed and presented in this paper. The first strategy is a non-linear Sliding Mode control scheme. The second strategy is an H-infinity control scheme. Both schemes are implemented in simulation, and both schemes show promising results for state convergence in the presence of variable cross-country time delays.