Hybrid Dynamic Systems refter to a class of systems which are composedof looth Continuous Variable Dynamic Subsystems (CVDS) and DiscreteEvent Dynamic Subsystems (DEDS). The conventional systems involvecontinous variable which can be describled by difference ordifferential equations. And DEDS is a finite state transition which isobserved and can be controlled by a language generated by the labeisthat are attached to each transition. However, in most applications, continuous variable subsystems and discrete event subsystems exist inthe same framework. Not only does the evolution of continuous variablesaffect the occurrence of discrete events, but the discrete events alsochange the performance of continuous variable system. In this paper, the author puts emphasis on the modelling and controlling of HDS. First, HDS is classified as tight HDS (THDS) andloose HDS (LHDS) on the basis of dynamic couple of discrete andcontinuous variables. The following research consists of two parts. Thefirst one analyses LHDS and the second one analyses THDS. In LHDS modelling framework, HDS consists of three parts: discretesupervisor, interface and continuous plant. The discussion emphasizesthe consistency of various model discretations of continous state spaceand the uncertainty of discretation. The author creats a new modelling framework ----Hybrid Petri Net (HPN) and uses it to analyse LHDS. In THDS, since dynamics of THDS is not easily decoupled, the modelling frameworkof THDS must be of relational model, rather than conventional input--output map. Finally, the property of dynamic system on finite module 3 ( F3 ) isanalysed, and some results of logic and synchronization are discussed.
修改评论