2003 年 16 巻 9 号 p. 461-467
The least mean square (LMS) algorithm has been widely used in adaptive signal processing applications, because of its simplicity and low complexity. This paper proposes a design method for an LMS-type algorithm which is robust in some sense and converges faster than the conventional LMS algorithm. The design problem is reduced to a semidefinite program which is an efficiently solvable optimization problem. Moreover the proposed design method is extended to the filtered-X LMS algorithm for active noise control, and an experimental result is provided to illustrate the effectiveness of the proposed method.