An Alternative Feedback Structure for the Adaptive Active Control of Periodic and Time-varying Periodic Disturbances

An alternative feedback structure, inspired from a speech coding structure, is introduced for the adaptive active control of periodic and time-varying periodic disturbances. The new structure is an Internal Model Control (IMC) structure. It separates the information and the processing pertaining to the plant from the information and the processing pertaining to the disturbance signal. A non-causal filter modelling the inverse of the secondary path is used, and the future samples of the disturbance signal are estimated by linear prediction (using an LMS algorithm). No filtering of a reference signal with the model of the secondary path is required to compute a filtered reference signal. In consequence, for the cases presented in the paper, the new feedback structure has a much faster convergence behavior and a better transient response than the classical FX-LMS algorithm with a feedback IMC structure, or the more recent and usually faster ANC-LMS algorithm with the same IMC structure. Simulation results with experimentally measured primary and secondary paths of a beam are presented, showing that either for multi-harmonic disturbance signals or sweeping sine wave disturbances, the structure presented in this paper greatly improves the convergence behavior of an adaptive feedback controller. 7 1998 Academic Press Limited