 |
Center for Nonlinear Studies |
Modern computer control systems are usually implemented by periodic sampling of sensors and zero-order hold of actuators due to the simplicity. However, many processes are triggered by specific events not time. Event-triggered sampling is an alternative way to perform A-to-D conversion in computer controlled systems. In this talk, a few simple scenarios are considered, in which we are able to obtain analytical results, to gain an insight into the event-triggered sampling. First, I will present the linear quadratic control problem for first order stochastic systems over a finite time horizon with limited sampling actions. The performance of three messaging policies for sensing combined with two hold circuits for actuation is compared based on optimization over the parameters of event detection and feedback control. The sampling rules include deterministic sampling (DS), level-crossing sampling (LCS) and optimal sampling (OS), and the hold circuits include zero order hold (ZOH) and generalized hold (GH). The general results are established that level-crossing sampling performs more effectively than deterministic sampling and generalized hold outperforms zero order hold. Next, I will demonstrate that event-triggered sampling outperforms periodic sampling in terms of mean-square state variations for two dimensional stochastic systems. The procedures are provided to design the optimal sampling period for periodic sampling and the optimal threshold for event triggered sampling.
Host: Anatoly Ziotnik