The patent badge is an abbreviated version of the USPTO patent document. The patent badge does contain a link to the full patent document.
The patent badge is an abbreviated version of the USPTO patent document. The patent badge covers the following: Patent number, Date patent was issued, Date patent was filed, Title of the patent, Applicant, Inventor, Assignee, Attorney firm, Primary examiner, Assistant examiner, CPCs, and Abstract. The patent badge does contain a link to the full patent document (in Adobe Acrobat format, aka pdf). To download or print any patent click here.
Patent No.:
Date of Patent:
Nov. 07, 1995
Filed:
Mar. 05, 1993
Per Enge, Los Altos, CA (US);
Dominic Farmer, Milpitas, CA (US);
Brian Westfall, Mountain View, CA (US);
Trimble Navigation Limited, Sunnyvale, CA (US);
Abstract
Method and apparatus for reducing or cancelling impulse noise from a signal containing noise. The desired noise-free signal is assumed to have a representative frequency .omega..sub.3, but may have a range of frequencies adjacent to this frequency, and is assumed to have substantially zero amplitude for all frequencies .omega.<.omega..sub.1 and/or for all frequencies .omega.>.omega..sub.2, where .omega..sub.1 <.omega..sub.3 <.omega..sub.2 or .omega..sub.1 <.omega..sub.2. An input (noisy) signal is filtered and analyzed in a narrow frequency region surrounding .omega.=.omega..sub.1 and/or a narrow frequency region surrounding .omega.=.omega..sub.2 to obtain one or two output signal components n.sub.1 (t) and/or n.sub.2 (t), respectively, that, ideally, contain no contribution from the desired signal. The input signal is also filtered and analyzed in a narrow frequency region surrounding .omega.=.omega..sub.3 to obtain an output signal s(t)+n.sub.3 (t) component including the desired signal s(t). A linear combination of signals, such as S(t)=s(t)+n.sub.3 (t).+-.[n.sub.1 (t)n.sub.2 (1)] .sup.1/2, or S(t)=s(t)+n.sub.3 (t)+.alpha.exp(-j.phi.) [n.sub.1 (t)n.sub.2 (1)].sup.1/2,or S(t)=s(t)+n.sub.3 (t).+-.n.sub.k (t) (k=1,2), or S(t)=s(t)+n.sub.3 (t)+exp(-j.phi.)n.sub.k (t), or S(t)=s(t)+n.sub.3 (t)+[exp(-j.psi.1)n.sub.1 (t)+exp(-j.psi.2) n.sub.2 (t)]/2, or S(t)=s(t)+n.sub.3 (t)+.alpha.[exp(-j.psi.) n.sub.1 (t)+exp(-j.psi.) n.sub.2 (t)]/2, is formed as a circuit output signal, and the .+-. sign and/or the multiplier .alpha. and/or the phase angles .phi., .psi.1, .psi.2 and .psi. are chosen to minimize the contribution of noise to S(t), according to a selected quantitative error measure. Three such error measures are the displacement of a signal from the median of a reference signal,the cumulative variation of a signal, and the least mean square value of a signal. Other combinations of the signals n.sub.1 (t) and n.sub.2 (t), replacing the (complex) arithmetic mean and geometric mean, may be used for the linear combination signal S(t). Methods for computation of the phase angles .phi. , .psi.1, .psi.2 and .psi. are also disclosed.