Analysis of artificial intelligence technology based on the requirements of collaborative robots through patent analysis

Abstract

Since the manufacturing market has been declining lately, many companies are paying greater attention to more-efficient production methods. The research on collaborative robots based on Artificial Intelligence (AI) technology, in particular, has become diversified as interest in “smart factories” has been receiving greater focus recently. This study analyzed the requirements and technology needed for using collaborative robots at industrial sites. Currently, Korea is one of the major countries in the global patent market along with China, the United States, Japan and Europe. Korea is the fourth largest country with 7.9% of the global patent market. Therefore, considering the amount of patent data that is registered in Korea, the data collection and analysis for this study are meaningful. Thus, we collected data from abstracts of patents based on the AI (G06Y) technology classification table provided by the Korea Intelligent Property Office (KIPO). In addition, we extracted words based on the reference of co-operating robot requirements by referring to KS B 7313 and conducted an analysis through a matrix that matched the requirements of AI technologies and collaborative robots. Furthermore, the statistical analysis program SPSS 24.0 was used to perform factor analysis and regression analysis. As a result, the 11 factors of the collaborative robot were grouped into two mega-factors. We were also able to identify the cause and effect relationships between the two factors. In addition, to improve the productivity and safety of collaborative robots, ergonomic design must be improved, and there are five factors that must be considered for the ergonomic design of collaborative robots. Furthermore, it was found that, in order to influence these five factors, it was necessary to recognize events of situation recognition among AI technologies, to recognize objects and places of visual recognition, to judge the map learning tree, to classify self-study, and to have deep trust network technology of deep learning.