Fault diagnostics in intelligent control systems of autonomous robots: An ontological approach
TL;DR
A new paradigm is identified for the development of a promising class of autonomous robotic systems with increased fault tolerance and, consequently, adaptability through the use of an ontological model used to establish cause-and-effect relationships between observed failures and the malfunctions that cause them.
Abstract
Objectives. The work set out to develop diagnostic tools and methods for autonomous robots for the rapid detection of faults based on the analysis of cause-and-effect relationships between observed manifestations and the sources of failures. Methods. The paper presents an analytical review of the current scientific and technical literature on the topic of the research, which includes ontology construction methods for developing a diagnostic model for autonomous robots, Boolean algebra methods for generating queries to the ontological diagnostic model, modifications to conditional and unconditional diagnostic methods for finding cause-and-effect relationships between failures and malfunctions based on the use of an ontological model, and machine experimentation methods for estimating the processing time of queries to the ontological diagnostic model. Results. The presented ontological model and methods for conditional and unconditional diagnostics is used to establish cause-and-effect relationships between observed failures and the malfunctions that cause them. Conclusions. Based on an ontological approach and taking into account the availability of advanced tool support, including ontology description languages, ontology modeling tools, and ontology-based inference and reasoning technologies, a new paradigm is identified for the development of a promising class of autonomous robotic systems with increased fault tolerance and, consequently, adaptability. The development of an autonomous robot presupposes the presence of an onboard intelligent control system built on a hierarchical principle that comprises executive, tactical, and strategic levels. The results of the ontological model provide an effective basis for activating and utilizing the intelligent control system’s rich algorithm databases at all levels of the control hierarchy to support increased adaptability.