OBSERVED VALUE ANALYSIS METHOD

摘要

PROBLEM TO BE SOLVED: To provide the high precision analyzed result by estimating the inverse matrix of an error covariance matrix from the variance of observation error so as to consider the timewise correlation of observation error based on integra tion information and estimating the quantity of state with the inverse matrix of error covariance matrix as weight. SOLUTION: An integration processing part 2 integrates instantaneous observed valuesβOt <(k)> ((k) is 1 to K; K kinds of observed values are used) inputted from an instantaneous observed value input terminal 1 and gives integrated observed valuesβt <(k)> to a calculation part 4. The calculation part 4 calculates error varianceδt<(k)> of observed values from the observed valuesβt <(k)> , applies observed valuesβt <(k)> to a state quantity estimating part 6 and applies the error varianceδt <2> (k) to an error covariance matrix inverse matrix estimating part 5. The estimating part 6 finds a state quantity vector Xo from the observed valuesβt <(k)> applied from the calculation part 4 and an inverse matrixΣ<-1> of an error covariance matrixΣof observed values applied from the estimating part 5 and finds a vector Xo at desired time (t) to provide the analyzed value from the vector Xo. Then, the result is applied to an output terminal 7.