Abstract:
Fault-tolerant control addresses the control of dynamical systems such that they remain functional after the occurrence of a fault. To allow the controller to compensate for a fault, the system must exhibit certain redundancies. Alternatively, one may relax performance requirements for the closedloop behaviour after the occurrence of a fault. To achieve fault tolerance for a hierarchical control architecture, a combination of both options appears to be advisable: on each individual level of the hierarchy, the controller may compensate the fault as far as possible, and then pass on responsibility to the next upper level. This approach, when further elaborated for discrete-event systems represented by formal languages, turns out to impose a hard lower-bound inclusion specification on the closed-loop behaviour. The present paper discusses the corresponding synthesis problem and presents a solution. © 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
