| 摘要 |
PROBLEM TO BE SOLVED: To enable an adaptive filter to easily change convergence characteristics of loads. SOLUTION: The adaptive filter comprises a calculation means for storing M-piece load values, performing the product-sum operation of M-piece latest input signals to the M-piece stored load values, deducting a result of product-sum operation from desired response signals, and calculating M-piece new load values based on the signals, the M-piece latest input signals, and the M-piece load values held by a load storage means. An input signal vector including the M-piece latest input signals as elements is referred to as xk, a load value vector having M-piece elements is referred to as wk, a residual signal for outputting 10 is referred to as ek, a gain parameter being a positive constant is referred to asα, a time delay parameter being a positive constant is referred to asβ, a time normalization parameter being a positive constant is referred to asγ, an integer for expressing a discrete time of digital signals is referred to as k, * is referred to as a complex conjugate, ¾¾xk¾¾ is referred to as a norm of the input signal vector xk. In such a state, a load calculating means sequentially updates the M-piece load values based on a relation of wk+1=wk+αekxk*/ä(kγ+β)¾¾xk¾¾^2}. COPYRIGHT: (C)2007,JPO&INPIT
|
