Robust control of hybrid systems with application to automotive powertrain system.

摘要

In this research, the problem of robust control of hybrid systems is investigated. The hybrid systems under study are characterized as systems whose inputs contain both analog variables and discrete actions; and whose behavior belongs to an uncertainty set which includes both metric modelling errors and discrete uncertainties.; A control structure is introduced which consists of robust controllers for analog control and supervisors for discrete events at lower levels, and managers to make switching decisions at higher levels. For a generic class of uncertain nonlinear hybrid systems, a design methodology for managers is developed based on system performance. Robust stability and performance of the closed-loop hybrid systems are established for systems subject to both analog modeling errors and structural uncertainties.; The hybrid design methods, as well as feedback linearization and nonlinear robust Hinfinity control, are applied to automotive powertrain systems. A feasible powertrain control strategy is developed.