Human Behavior Understanding Based On Expert System In Intelligent Space

With the development of society and the advancement of technology, the demand for robots is no longer limited to teaching and entertainments. How to provide autonomous and intelligent service for people in the daily life has become a topic in the research field of robots. Human behavior recognition and understanding is the premise for robots to provide intelligent services, and it is also the assurance to assist and improve the quality of life for the elderly.In this paper, in service robotic lab of Shandong University, we utilize intelligent space technology and expert system to complete the recognition and understanding of some basic daily activities, at the same time, finish the forecast for the potential daily behavior for the elderly successfully. The main contributions of this thesis are shown as follows:Firstly, a scheme based on context-aware system in intelligent space and expert system is constructed to recognize and understand human behavior. Then context information related to human behavior is obtained by constructing context-aware system and designing a heterogeneous network model. Besides, CLIPS expert system is constructed to reason human behavior.Secondly, According to the characteristics of the knowledge used for behavior understanding, confidence is selected to measure the uncertainty of knowledge. The confidence of all the information is obtained by many acquisition experiments of context information. Additional, a reasoning method based on confidence is proposed. Then, production representation and object-attribute-value triple are adopted to represent behavioral rules and facts respectively. At last, rule base and fact base are designed to save rules and facts for behavior understanding.Thirdly, because of the limitation of the fixed number of rules for behavior understanding, rough set theory is adopted to realize the expansion of the rule base automatically. New simple rules are generated finally by using the algorithm of attribute reduction and attribute value reduction. Through knowledge reduction, the rule base is expanded in a certain extent, which improves the accuracy of the behavior understanding.Finally, all functional modules are integrated to form a software system for behavior understanding, and simulation experiment is carried out to recognize people’s daily behavior.