Computational complexity reduction in multi-processor implementation of recursive digital filters by similarity transformation
K. Negoto, Shigenori Tomiyama, Tsuyoshi Takebe · 2002
Efficient computation method for recursive digital filtering is studied in multi-processor environment. The method reduces computational complexity by applying similarity transformation to the state update matrix in the block state-space representation. In implementing a cascade of 2nd order sections, it is found that the whole system is most advantageous if described as a single stage.