Control Design And Analysis For Systems Subject To Communication Or Computational Constraints

The rapid revolution of the communication and computational technologies enable a new trend for the development of control systems,where many emerging systems of new form such as the Internet of Things,cyber-physical systems,all depend on complex communication networks and various embedded computing devices to close the control system.Here,the wide applications of the communication and computational technologies have greatly broadened the applicability of control systems,while at the same time the constraints induced by communication and computation are also introduced to control systems which require new approaches to deal with.This thesis investigates the design and analysis of two special classes of communication/computation constrained systems.The mathematical definition of the problems is given,solutions are proposed,and the effectiveness are verified by numerical examples.The works contained in this thesis include,1.The communication constraint problem of multi-delay caused by multi-packet transmission is investigated.The wide use of the wireless communication makes it popular for the control systems to use multiple sensors,while the latter causes the problem of “multi-packet transmission”,i.e.,the system output is transmitted via multiple independent data packets,which then consequently causes the partially inconsistent delay for the system output,a new challenge for the design.The mathematical definition of the problem is given,a packet-based active compensation scheme is proposed based on the predictive control idea,the closed-loop stability is analyzed,and the effectiveness of the approach is finally verified using numerical and TrueTime based simulations.2.The computational constraint problem of lacking computational resources due to the sharing of these resources is investigated,which is common in systems such as the Internet of Things where shared embedded computing devices are broadly used.A fullresource predictive control method is proposed following the co-design idea,which optimizes the system performance at no extra cost of the computational resources.The closed-loop stability is analyzed with discussions on the expendability of the approach,and the effectiveness of the approach is verified numerically.