A Spectrum-based Filter Design for Periodic Control of Systems with Time Delay

Abstract

A fully analytical controller design is proposed to tackle a periodic control problem for stable linear systems with an input delay. Applying the internal model control scheme, the controller design reduces to designing a filter, which is done through the placement of poles and zeros. The zeros are placed to compensate for the harmonics and to achieve the desired degree of properness for the filter. For placing the poles, a quasi-optimal procedure is proposed utilizing the standard LQR method. Given the high-dimensionality of the filter due to targeting a large number of harmonics, the design, as well as controller implementation, is performed over a state-space representation. A thorough experimental case study is included to demonstrate both the practical feasibility and effectiveness of the proposed control design. The experimental validation is performed on a physical system, the goal of which is to reject periodic vibrations acting on a mass-spring-damper setup where the sensor and the actuator are non-collocated.