Adaptive linear estimation of dynamic measurement error

Abstract

The problem of the dynamic measurement error estimation and compensation is considered. This type of error is determined by two components. The first one is due to dynamic properties (inertia) of a sensor. The second one is due to the presence of an additive noise at the sensor output. An approach to estimate and reduce the dynamic measurement error based on the signals adaptive linear prediction or adaptive line enhancement principle is proposed. The approach consists in generating a dynamic measurement error estimation signal based on comparing a delayed copy of the recovered signal with the recovered signal passed through an adaptive non-recursive filter with a linear phase characteristic. The structure of a measuring system with an adaptive linear estimator of the dynamic measurement error based on this approach has been developed. A computer simulation of the proposed measuring system for the second-order sensor is carried out. Optimal (in the sense of the mean squared deviation of the dynamic error) orders of the restoring adaptive filter are obtained in the presence of additive harmonic noise of variable frequency at the sensor output. The properties of the proposed measuring system with the dynamic measurement error estimator adaptive to the noise parameter are demonstrated. The application field of the results obtained is the measurement data processing of fast-changing processes (including real-time mode), when the component of the dynamic measurement error, caused by dynamic properties (inertia) of the sensor, as well as additive noises at its output, is dominant. The solution of such a problem is relevant, for example, when processing the results of ground tests of space technology.