抄録
The Filtered-x LMS algorithm used from the former as an adaptive algorithm of active noise control is simple, and extremely stable. However, identification of a secondary path characteristic is required prior, and control becomes impossible when the error of the identified characteristic and an actual acoustic characteristic becomes large. This poses a practical problem. Authors are aiming at solution by the technique called Direct Adaptive Algorithm (DAA). In DAA, carrying out sound control and secondary path characteristic identification are carried out simultaneously. Prior acoustic characteristic identification is less necessary by this. Since it identifies accommodative also about the problem of a modeling error, it is solved. Since a secondary path characteristic is identified accommodative, ANC by a direct adaptive algorithm can make control follow movement of an evaluating point. However, since adaptation speed is not so high-speed, the control effect deteriorates during movement of an evaluating point. In this report, in order not to degrade the control effect also while moving, when an evaluating point moves in ANC, the method of taking time delay into consideration to updating of an adaptation filter is proposed, and the validity of the method is confirmed by a numerical simulation.
