D.A. Masterenko

Results of measurements in some practical cases are highly quantized. The random scatter caused by both observed process and random error turns out to be as high as quantization step size of the measuring device. Moreover, alterations of observed value are also comparable to quantization step size. This situation is quite common for most precise measurements. In order to statistically process highly quantized observations the author proposed and considered special statistical estimates, called Pittman-type estimates. This kind of estimates made it possible to obtain more precise results compared to traditional estimates. Pittman-type estimations calculation can be realized using numerical integrating methods, up-to-date computers and software. Still, the more observations are made, the higher the computational complexity is. Moreover, the data are gathered serially in real-time measuring systems, thus the volume of data increases rapidly and new pieces of data are request recalculation of estimate. The author presents recurrent modification of Pittman-type estimates, lowering the volume of computations for real-time systems. The results of recurrent and non-recurrent estimates calculations are presented.