Krachkovsky M. Information technology of situational control with unsupervised learning, based on cognitive psychology models

Object: machine learning processes. Subject: unsupervised learning model in situational control systems. Methods: methods of control theory (situational control), methods of artificial intelligence (fuzzy sets and fuzzy logic theories), methods of machine learning (reinforcement learning). Research purpose: expansion of the scope of the effective use of production systems operating in open environments, through the use of machine learning in situational control technologies. Theoretical and practical results: developed in the work method, models and information technology can solve a wide class of problems associated with control of robotic systems in open environment, allowing the operational machine learning. Scientific novelty: method of machine learning of situational control systems is proposed for the first time, which differs from the prior ones that it uses generalized and formalized learning models, which are investigated and proposed in cognitive psychology; machine learning method gained further development by taking into account data from cognitive psychology about influence of context, situation and motive on learning mechanisms; information technology of situational control is improved by modifying the model of context-motivated chained situational control with operational learning mechanism.