Feedforward adaptive noise control with multivariable gradient lattice filters

This work presents a novel feedforward adaptive noise control (ANC) algorit hm based on multivariable gradient lattice filters to control acoustic nois e or vibration globally. In addition, a gradient-based lattice for AR and F IR models and its transpose lattice for the multivariable ANC algorithm are derived. The filter has different forward and backward reflection coeffici ent matrices to provide a faster convergence than the gradient lattice algo rithm when using the same reflection coefficient matrices. Experimental res ults demonstrate the effectiveness of the proposed algorithm in controlling broadband acoustic noise in an enclosure.