Research And Application On Affordance Model Of Cognitive Robot

With the continuous development of artificial intelligence, robots close to humans fromall aspects and in some respects beyond the human. But in the face of complex and volatileenvironment, the humans ability to perform tasks is far better than robots. So people began tostudy cognitive robot. Bio is the best robot imitation objects, establish bionic model is animportant way to enhance robot cognitive abilities. Affordance (action potential) theory ofcognitive robotics, proposed by psychologist J.J. Gibson, provides a new way to study.However, the current cognitive robot research is still in the initial stages and the effect ofperforming tasks is not ideal. Therefore, the research of cognitive robotics is meaningful andhas broad application value.This paper includes the following several aspects:First, illustrate the research background, signification, development process and researchstatus.Second, describes the affordance (potential action) theory, related work and five mainformal systems and two models.Third, introduce the concept of analytic functions according to Visual Attentionmechanism and established the affordance model of cognitive robot based on the analysisfunction.Forth, introduces the hierarchical reinforcement learning theory into affordance model,and use rescue robot experiment to describe and validate the model.Fifth, summarize the two affordance models of cognitive robot and experimentspresented in the paper, those experiments have achieved good result, verify those models arepractical.