Method for validating the value of a parameter
First Claim
Patent Images
1. A signal validation method for generating a reliable signal from a plurality of redundant received signals, comprising the following steps:
- simultaneously measuring samples of each of said received signals;
memorizing the values of the last N samples (N≧
2) of each received signal;
determining a plurality of model functions corresponding to the N memorized samples for each of said received signals, said model functions being defined as a plurality of parameters representing an approximate function for said N memorized samples;
calculating a dispersion coefficient representing the difference between each sample of the N memorized samples and a sample, determined by the model functions;
validating said model functions by comparing said dispersion coefficient with a predetermined dispersion threshold, said model function being valid if said dispersion coefficient is smaller than said predetermined dispersion threshold; and
generating said reliable signal by determining a value for a sample of said reliable signal as a function of at least some of said valid model functions.
1 Assignment
0 Petitions
Accused Products
Abstract
The method according to the invention consists of sampling, at successive times separated by constant or non-constant time intervals, signals supplied by redundant channels measuring or calculating the parameter and validating a value of said parameter, as a function of a sample of N values of each channel, in which N≧2. In a preferred manner, each sample of N values is modeled by a straight line using the method of least squares and the validated value of the parameter is worked out as a function of the inter-coherent models.
-
Citations
11 Claims
-
1. A signal validation method for generating a reliable signal from a plurality of redundant received signals, comprising the following steps:
-
simultaneously measuring samples of each of said received signals; memorizing the values of the last N samples (N≧
2) of each received signal;determining a plurality of model functions corresponding to the N memorized samples for each of said received signals, said model functions being defined as a plurality of parameters representing an approximate function for said N memorized samples; calculating a dispersion coefficient representing the difference between each sample of the N memorized samples and a sample, determined by the model functions; validating said model functions by comparing said dispersion coefficient with a predetermined dispersion threshold, said model function being valid if said dispersion coefficient is smaller than said predetermined dispersion threshold; and generating said reliable signal by determining a value for a sample of said reliable signal as a function of at least some of said valid model functions. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11)
-
Specification