A novel method of dynamic correction in the time domain

Abstract

The dynamic error of measured signals is sometimes unacceptably large. If the dynamic properties of the measurement system are known, the true physical signal may to some extent be re-constructed. With a parametrized characterization of the system and sampled signals, time-domain digital filters may be utilized for correction. In the present work a general method for synthesizing such correction filters is developed. It maps the dynamic parameters of the measurement system directly on to the filter coefficients and utilizes time reversed filtering. This avoids commonly used numerical optimization in the filter synthesis. The method of correction is simple with absolute repeatability and stability, and results in a low residual error. Explicit criteria to control both the horizontal (time) and vertical (amplitude) discretization errors are presented in terms of the utilization of bandwidth and noise gain, respectively. To evaluate how close to optimal the correction is, these errors are also formulated in relation to the signal-to-noise ratio of the original measurement system. For purposes of illustration, typical mechanical and piezo-electric transducer systems for measuring force, pressure or acceleration are simulated and dynamically corrected with such dedicated digital filters.