Standing-in for a Hammer: An Incremental Learning Based Approach to Tool Substitution (Part 1)
Consider a robot performing a task that involves tool use. For instance, a service robot is asked to serve drinks on a tray, but it realizes that the tray is broken; such mishaps in day-to-day activities are common. In situations like these, an effective way for a robot would be to find an alternative as humans do, for example, use an eating plate for serving, rather than wait until a tray becomes available. Since an availability of a tool can not be assumed, this skill is significant when operating in a dynamic, uncertain environment because it will allow a robot to adapt to unforeseen situations to a degree.The question is how can a robot determine which object from the environment to choose? In my doctoral work, I would like to propose an approach based on an incremental learning to address the problem of substitution. The proposed approach takes cues from Gibson's theory of affordances and cognitive linguistic theory of image schema to determine a substitute. In this presentation, I would like to discuss the approach and present the preliminary results.