APPLICATION OF A MODIFIED KALMAN FILTER FOR THE TRANSIT-TIME CORRELATION

A known procedure for non-contact speed measurement of objects with random surface characteristics is the transit rime con elation. Nevertheless, it has hardly been used in industrial practice. One of the reasons is to be found in stochastic measuring errors. Their occurence probability depends upon the sampled properties of the object as well as upon the dynamic systems behaviour of the applied sensors, and is difficult to be determined. This paper deals with a modified Kalman-filter, which distinctly decreases the influence of measuring errors on the variance by including a priori known informations about the dynamic behaviour of the measured object in the velocity estimation. Moreover, it describes the basic requirements of the correlator with regard to a succesfull application of the optimal filter and the available improvements of the resulting variance.