Intelligent state assessment of complex autonomous objects based on wavelet analysis

Abstract

The evolutionary development of complex autonomous technical objects (CATOs), which include bodynets, unmanned vehicles , aircraft, and other robotic systems , is characterized by increased requirements for the quality of their functioning, security, and reliability under the influence of various kinds of destabilizing factors. This determines the importance of the problem of assessing their technical state, including monitoring degradation (aging) and supporting the dynamic adaptation of CATOs. To solve this problem, the article proposes a new method for intelligent assessment of the CATO technical state. The method is based on the representation of knowledge about the results of interval estimation of controlled parameters in the knowledge base of the system for assessing the CATO technical state. In this case, the process of estimating the controlled parameters is based on the application of wavelet analysis . The article discusses the architecture and implementation issues of the intelligent system for assessing the CATO technical state. A special place in this system is occupied by the knowledge base containing information about the emergency and normal states of controlled parameters. An experimental evaluation of the proposed assessment method showed that the joint use of knowledge representation processes in the knowledge base and wavelet analysis for the formation of CATO operability regions increases the accuracy and credibility of the state identification results. In addition, this approach expands the possibilities of applying technical means of control and diagnostics concerning evolving CATO. • A new method for intelligent assessing the state of complex autonomous technical objects (CATOs) based on wavelet analysis. • A mathematical and algorithmic foundations for assessing the technical state of CATOs. • An architecture and implementation of the intelligent system for assessing the CATO technical state. • An evaluation of the proposed method showed the increased accuracy and credibility of the state identification results. • Comparison the proposed approach with the existing methods based on Local Spline Models and Least Squares.