Reduced-Order State Observers Design for Single-Channel Electromechanical Control Plant under Conditions of Parametric Uncertainty

Abstract

The observation problem for a single-channel electromechanical control plant operating under conditions of significant parametric uncertainty is considered. Decomposition procedures for the synthesis of reduced-order state observers with piecewise linear control actions have been developed. The proposed approach provides for a given time the specified estimation accuracy of the unmeasured state variables of the mechanical subsystem according to the information received from the sensors of the state variables of the electric actuator. The proposed approach is fundamentally different from traditional methods, where the main equations for constructing a reduced-order observer are equations describing the dynamics of unmeasured variables. In this paper, the observer is constructed as a copy of the differential equations of the measured variables of the electrical subsystem with certain parameters. In these equations, unmeasured state variables are treated as undefined inputs, for which the evaluative signals are the control actions of the observer, which are limited in magnitude. Due to such an organization of the observation subsystem, the problem of identifying uncertain mass-inertial parameters of the mechanical subsystem does not require a solution. This greatly simplifies the adjustment of the observer, eliminates bursts, and ensures the robustness of the estimated signals.