Inverse Model Approach to Disturbance Rejection Problem

Disturbance rejection and output tracking problem in linear multivariable control systems is considered. Using inverse dynamic models approach, controller structures including disturbance observer and compensator are proposed for disturbance estimation and compensation. The method for inverse dynamic models synthesis based on unknown input observer theory is justified. The properties of closed-loop control system with inverse model-based controller have been investigated for the purpose of attainable disturbance rejection accuracy assessment. Moreover, nonsingularity conditions of controlled plant structure are established, allowing eliminating disturbance estimates from decoupling compensator equation. For this case, when such conditions are violated, the realizable form of disturbance decoupling compensator includes internal dynamic filter with small time constant being proposed. It was shown that slow motion in two-time-scale system coincides with the processes in the system with ideal compensator, if the fast motion in closed-loop system is stable. A generalization of inverse model method for systems with variable structure is also considered. Sliding mode equivalence principle is proposed to design non-stationary sliding surfaces with integral parts, in that, inverse modelbased equivalent control ensures mismatched disturbances compensation. Variable structure disturbance observer and compensator are also developed and sliding mode conditions for proposed control scheme are established. 130Peculiarities of applying inverse model approach for discrete-time systems are also considered and procedures for structural–parametric synthesis of discrete disturbance observers and compensators are suggested.

Перевод пока недоступен