Open AccessModel and method development for determining the completeness of information for remote detection of landmarks for autonomous mobile robots
Author(s) -
Oleksandr Poliarus,
Andrii Lebedynskyi,
Yevhenii Chepusenko,
Nina Lyubymova
Publication year - 2021
Publication title -
technology audit and production reserves
Language(s) - English
Resource type - Journals
eISSN - 2706-5448
pISSN - 2664-9969
DOI - 10.15587/2706-5448.2021.244663
Subject(s) - completeness (order theory) , a priori and a posteriori , landmark , robot , computer science , mobile robot , artificial intelligence , data mining , computer vision , mathematics , mathematical analysis , philosophy , epistemology
The object of research is the completeness of information for making a navigation decision by an autonomous mobile robot when it performs a task in an unfamiliar area without GPS. It is difficult to identify a landmark in the absence and abundance of information. One of the most problematic places is the mathematical description of the criterion according to which an autonomous robot makes a decision about the completeness of information. The paper substantiates a model and method for determining the completeness of information by a robot equipped with several landmarks detection tools operating on different physical principles. It is shown that the implementation of the method requires a priori information on the probability of detecting various landmarks by passive and active means against a continuous and discontinuous background at different illumination of objects, in day and night conditions under different weather conditions. The values of the probability of detecting a specific type of landmark obtained in such studies serve as the basis for constructing an information cadastre for a job performing tasks on the ground. Three formulas are proposed for determining the coefficient of completeness of information, taking into account a priori and a posteriori inventories, and recommended areas of application. The value of this coefficient depends on the threshold level of the probability of detecting a landmark. The reliability of a decision made by a robot is greatest when it is made under conditions of a certain level of completeness of information. The proposed method can be used for other technical objects from which the measurement information is received. Compared with the known methods, it expands the boundaries of application and reveals the possibility of assessing the completeness of information in constantly changing conditions. Along with a change in these conditions, the characteristics of the completeness of information also change. The coefficient of completeness of information can approach unity even in the absence of separate means of detecting landmarks, and then the method makes it possible to assess the need for their use in the given conditions.