Open AccessON ALGORITHM FOR DETECTING CHANGES IN THE OBJECT MOTION MODE
Author(s) -
Aleksey V. Golubkov
Publication year - 2021
Publication title -
avtomatizaciâ processov upravleniâ
Language(s) - English
Resource type - Journals
ISSN - 1991-2927
DOI - 10.35752/1991-2927-2021-3-65-48-55
Subject(s) - kalman filter , trajectory , algorithm , motion (physics) , mode (computer interface) , computer science , a priori and a posteriori , noise (video) , state space representation , gaussian , object (grammar) , extended kalman filter , control theory (sociology) , mathematics , computer vision , artificial intelligence , image (mathematics) , philosophy , physics , control (management) , epistemology , quantum mechanics , astronomy , operating system
The article deals with the solution to the problem of determining the motion mode of an object along a complex trajectory. A hybrid stochastic model is used to describe a complex trajectory. The solution of the problem is based on the application of a sequential decision rule about the choice at an unknown time of the hypothesis about the current mode of motion, with a limited size of the bank of competing Kalman filters. An algorithm is constructed for calculating the average size of the Kalman filter bank in the case of M-possible motion modes. The algorithm is developed in a general form, therefore, it can be used not only for the four types of object motion models considered in this paper, but also for any linear discrete-time models with Gaussian noise presented by equations in the state space. The algorithm for a priori estimation of the average size of a bank of competing Kalman filters for M-possible modes of motion is implemented in MATLAB, the results of computer simulation are presented.