Decoupling Controller with Disturbance Observer for LTI MIMO Systems

Abstract

In this paper, a decoupling multivariable control strategy for linear time‐invariant (LTI) multi‐input/multi‐output (MIMO) systems is proposed. The strategy includes a multivariable disturbance observer (MDOB) and a decoupling controller. This MDOB is introduced to improve the system performances when the system encounters severe external disturbances. H2 optimal scheme is utilized to design the MDOB filter. The controller is developed based on an inverse control method, through which the design process can be simplified. Simulation results certify the effectiveness of the proposed control strategy.

A design of disturbance observer in frequency domain for robust control of time delay system

“…We selected sensitivity weight as W S = 1∕s 2 so that the system can remove ramp typed disturbance (q = 1) and therefore we can attain the Specification 2. However, since it does not satisfy the preconditions for standard H ∞ control problem, we replace it by W C = 1∕(s + 𝜆) 2 for small positive number 𝜆 = 0.01.…”

“…Step 3: Solution of optimization problem. We select the pseudo plant as G = 1∕(s + 𝛿) 2 , where 𝛿 could be arbitrarily selected for example, 𝛿 = 10, because the value of 𝛿 > 0 does not affect on the result. The order of augmented plant of the…”

“…After this preparation, we use the solution procedure of standard H ∞ problem to solve (31) by using hinfopt function of MATLAB Robust Control Toolbox. The optimized pseudo controller K * (s) was obtained at 𝛾 = 3.8825 as K * = 312.4738(s + 1.281)(s + 10) 2 (s + 103.1)(s + 0.01) 2 , where double zeros −10 are from double poles of G and double poles −0.01 are from 𝜆 of W S . The order of the augmented plant is fourth while K * is of third and its relative order is zero.…”

“…Disturbance observer based control (DOBC) method originally proposed in Reference 1 has been widely applied in high performance robust control due to its high disturbance suppression ability and simple structure. [2][3][4] As the performances in DOBC system highly depend on a low-pass filter, called Q-filter, design of the filter has been essential in DOBC system design and analysis, and has attracted much attention in academic society. [5][6][7][8] In general, the disturbance suppression performance concerns with the Q-filter's bandwidth.…”

Inverse‐compensation‐based DETC for nonlinear hysteresis system with disturbance

“…In ADRC, an extended state observer is used to estimate disturbances and then compensate for them in a feedforward way [44][45][46]. In DOBC, a linear/nonlinear disturbance observer is designed to obtain disturbance information and then attenuate the influence of the disturbance using proper feedback control [47][48][49][50]. In the above-mentioned methods, the disturbance is always eliminated directly as a negative element.…”