Designing a Support System for Action Rule Extraction in Script-Based Robot Action Planning

In script-based robot action planning, a script representing procedural knowledge is used as a form of a task model. To make much of efficiency while dealing with HRI, this research suggests that action rules have to be found using the script as a source. These action rules are used as an analyzing tool for the script, and the analysis results are then utilized to keep the script in good working order. This research defines three types of action rules. The data mining method is adopted to discover the action rule. This research proposes two approaches that improve performance based on action rules when utilizing a script. Given that a computer system that supports all of the work related to managing rules is developed throughout the process, it would show the progress systematically and facilitate all types of work.