Modeling and H∞ optimal control of active noise cancelling headphones

This paper presents the preliminary results of an ongoing study to maximize the attenuation level and bandwidth of active noise cancelling (ANC) headphones through the use of advanced system identification methods and feedback control design tools. After conducting an identification experiment on the physical system, the collected data were used to estimate a headphone model. The “best” model was obtained by comparing the results of several identification methods, including prediction error methods based on ARX and state space models as well as nonparametric methods. H_∞ optimal control and μ-synthesis techniques are shown to be ideally suited to the design of high performance robust active noise cancelling headphones. By choosing different weighting functions, performance limits (in terms of noise attenuation level and attenuation bandwidth) achievable by the system for various uncertainty levels were explored