Decentralized Tracking Of Interconnected Systems With Dynamic Input And Output Interactions

In the modern society, a variety of complex interconnected systems do commonly exist,such as: power systems, energy systems, aerospace systems, and so on. Control research hasbeen expanded from the initial system stability to tracking ability, and tracking studies havealso experienced a progressive leap from the basic asymptotic tracking to practical tracking.The decentralized tracking problems on a class of nonlinear interconnected systems are ex-plored, which is designing the control criteria with decentralized control technique, for anygiven satisfying certain assumptions of references and various subsystems with differentcharacteristics. Under the premise of ensuring the stability of entire system, we can achievethat the output of each nonlinear subsystem tracks any given reference signal asymptoticallyor practically, eliminating the in?uence of various disturbances including external perturba-tion and the interactions from other subsystems, etc. Also, the tracking errors converge tozero or a small set around the origin definitively. In this thesis, the tracking problems onseveral kinds of nonlinear interconnected systems, with the dynamic interactions and un-modelled dynamics depending on both the system inputs and outputs, will be studied withthe tool of decentralized adaptive backstepping approach, for the references meeting differ-ent conditions and the stochastic noise availability or not. Now, the main achievements willbe introduced concretely.Firstly, the asymptotical tracking on a class of nonlinear interconnected systems withthe dynamic interactions and unmodelled dynamics is explored. Decentralized adaptivebackstepping controllers and parameter update laws can make sure that outputs track anygiven reference signals asymptotically, and all the signals in the interconnected systemsare globally uniformly bounded. In order to extend the references range of the abovementioned system, practical tracking problem is considered with the relaxed referencerestrictions. Not only it enlarges the scope of reference signals, but also enhances thepractical applications of investigated systems. Besides the common system models,stochastic noise as the important disturbance to interconnected systems will also be studied.The output tracking on the stochastic nonlinear interconnected systems will be discussedwith and without parameter uncertainty. Decentralized adaptive backstepping controllers and parameter update laws will make sure the output can globally track any given referencesignals in the way of stochastic disturbance attenuation. Meanwhile, in order to enhance thestochastic system's tracking ability, the practical output tracking will be considered with thefirst-order differentiable references. Similarly, decentralized adaptive backstepping lawswill ensure the closed-loop output can globally track any given reference signals in the wayof stochastic disturbance attenuation, with the tracking errors converging to a small residualset around the origin in probability.