Bandwidth constraints of disturbance observer in the presence of real parametric uncertainties
Control systems based on disturbance observer (DOB) require a low pass filter (LPF) to avoid the algebraic loop and improperness in the inner-loop. Bandwidth of the LPF of DOB is desired to be as high as possible to improve the performance of a system. Two main constraints, which are noise and robustness of a system, determine the bounds on the DOB's bandwidth. Although the bandwidth constraint due to noise, which causes an upper bound on the bandwidth of DOB, directly depends on the sampling rate and measurement plants and methodology, the constraint due to robustness is affected significantly by robustness analysis methods. Conventional robustness analysis methods limit the bandwidth of DOB unnecessarily due to conservatism. This paper proposes a new robust stability analysis tool for control systems based on DOB in the presence of real parametric uncertainties. The proposed method shows that there is only a lower bound on the bandwidth of DOB to obtain robust stability, and the stability margin of a system increases as the bandwidth of DOB is increased. Consequently, the robustness of a system improves as the bandwidth of DOB is increased in the presence of real parametric uncertainties.