Метод определения количества бездефектных испытаний программного обеспечения автоматизированной системы подготовки данных полета беспилотных летательных аппаратов

Abstract

Calculating the reliability indicator of software for an automated system of preparing unmanned aerial vehicles flight data is special because of the variety of permitted points of the vehicles departure, one of which cannot be determined in advance. Since data quality control requires significant time resources, it is necessary to define a criterion according to which the entire amount of data is either accepted or rejected, or tests continue. Existing methods for assessing the software reliability index have shortcomings that prevent methods to be practically used in assessing the flight data quality. In this regard, we developed a new method based on monitoring the selection of test variants of input data from the permissible region using the results of syntactic and semantic control of the minimum required amount of test variants. The method assesses the quality indicator of the flight data of unmanned aerial vehicles with a given probability of an error of the second kind and is easily implemented in practice. For its implementation, we give the concept of an auxiliary sampling plan and prove a lemma on the uniqueness of this plan and theorems on the existence of quasi-optimal sampling plans according to the criterion of the minimum dead zone for making a positive decision with two and one negative test outcomes, as well as a defect-free plan. We also developed decisive rules for making a positive or negative decision based on the results of software tests.