专利名称：ADAPTIVE FILTER发明人：WEIGL, Konrad Michael申请号：EP93919204.0申请日：19930827公开号：EP0612447A1公开日：19940831
摘要： Adaptive filter for the approximation of a target function F, stored in a functionspace containing a set of n elementary filters gb, n being an integer 1. The elementaryfilters cover a sub-space in the space of functions, the parameters of the elementaryfilters being variable and the filter performing the following operations: calculating thenormal projections of the set function F on each elementary filter and multiplying thenormal projection quantities thus obtained Aa, respectively by the metric tensorcontravariant gab formed by the set of elementary filters, for the projection of the setfunction F on the sub-space covered, of which the difference D relative to the set functionF is determined as a function of the elementary filters and metric sensor, according toformula (I) so as to minimize the difference D by variation of the parameters of the filterelemen tary by the usual methods of deterministic calculation (e.g., gradient reductionmethod \") or stochastic (eg\" simulated annealing \approximate function being obtained by superposition of the elementary filters settledby minimizing, according to formula (II).
申请人：WEIGL, Konrad Michael
地址：Friedenspromenade 1a D-81827 München DE
国籍：DE
代理机构：Grünecker, Kinkeldey, Stockmair & Schwanhäusser Anwaltssozietät
更多信息请下载全文后查看
因篇幅问题不能全部显示，请点此查看更多更全内容