The Stochastic Domain Design of a Real-Time Controller for an Active Noise Control Headrest based on Finite Element Analysis

摘要

Designing an active noise controller in the stochastic domain has been proposed previously as an approach to account for the possible discrepancies between the actual and the nominal plant frequency responses. In this method, the probability distributions of the possible plant responses are used as a priori information and the controller is designed to minimize the expected value of the signal power measured by the error sensors with all possible plant responses. In active noise control, this approach has only been used to design the controller frequency response, but further practical issues were not considered, such as the controller causality and robustness as well as the real-time implementation of the controller through digital filtering. In the present work, the stochastic domain approach was implemented in the design of a real-time multi-channel active noise control headrest in an automobile environment where an Internal Model Control structure is adopted, FIR filters are used to construct the controller, and robust stability is considered. Finite Element Analysis was used to calculate the plant responses for the possible head locations and gestures in the actual operating condition of the system. The performance of the designed controller has also demonstrated in Finite Element simulations.