Invariants in problems of pattern recognition and control of states of complex objects

Abstract

Objective. The aim of the study is to build emulators of the main sensors of a small spacecraft (SC).Method. The quality of recovery is determined by the standard deviation (RMS) of the data from the standards. The study is based on the application of the dynamic programming method.Result. A review of works in the field of application of stable characteristics - invariants, in the applied areas of pattern recognition and state control of complex technical systems is carried out. A method aimed at identifying failures and building small spacecraft sensor emulators based on the analysis of stable indicators: correlation and covariance coefficients is proposed and tested. The proposed approach makes it possible to increase the overall reliability of the information-measuring system of a small spacecraft (SC). In the future, to emulate a failed sensor, it is planned to use multiple correlation of time sequences.Conclusion. In the general case, recognition is reduced to comparing the graph structures of objects to the correct matching of hierarchical image models based on various invariants. Control of the states of complex systems is based on the use of stable characteristics that act as invariant or subinvariant descriptors. Invariants in various applications help to recognize the current situation with a sufficient degree of accuracy, usually reducing the dimension of the problem.

Keywords

• hierarchy
• image
• graph
• isomorphism
• invariant
• tuple
• criterion
• complex object
• distance measurement
• pyramid of visibility
• control
• state