Active control of time-varying broadband noise using online system identification with parallel fast-array recursive least squares filters

For broadband active noise control applications with rapidly changing transfer paths, it is desirable to find algorithms with rapid convergence, fast tracking performance, and low computational cost. Recently, a numerically stable algorithm has been presented: a convex mixing fast-array RLS filter. The algorithm exhibits the fast convergence, tracking properties and the linear calculation complexity of the fast-array RLS sliding window filter with numerical stability. The algorithm assumes that deviations of the secondary path from the nominal value are small. If these deviations become significant then the model of the secondary path has to be updated in accordance with the real physical situation. In this paper we present results of a version of the algorithm with on-line adjustment of the secondary path model, using two mixing fast-array RLS filters, in which the first mixing filter minimizes a modified error signal while the second mixing filter minimizes the modeling error of the secondary path.