When we talk about objects in the environment, we indicate to an observer which object is currently under consideration using pointing gesture and the reference terms such as“this”and“that”. These behaviors play an important role in a natural interaction between us. It is because reference terms and pointing gesture can quickly inform the observer of an indicated object location. In this research we propose the 3-layered draw-attention model for communication robots with pointing gesture and verbal cues. The 3-layered model we propose consists of 3 sub-models, Reference Term Model (RTM), Limit Distance Model (LDM) and Object Property Model (OPM) . RTM is the model that decide the appropriate reference term using functions that is constructed by an analysis of human behavior. LDM is the model that decide whether or not the model use the object property with reference term. OPM is the model to decide the appropriate property of the indicated object comparing object properties each other. The 3-layered model was implemented as attention drawing system in the communication robot named, Robovie, and was confirmed its effectiveness.
