Autonomous object modeling and exploiting: A new approach based on affordances from continual interaction with environment

We present an architecture for self-motivated agents to generate behaviors in an environment that is continuous, both in space and time, through a continual interaction process. The long-term goal is to design agents that construct their own knowledge of objects and space through experience of the environment, rather than exploiting pre-coded knowledge. The agent exploits this constructed knowledge to exhibit behaviors satisfying its self-motivated principles, based on valences attributed to interactions, that specify inborn behavioral preferences. Over time, the agent learns the relation between its perception of objects and the interactions that they afford, in the form of data structures, called signatures of interaction. The agent keeps track of enacted interactions in a spatial memory in which it can use signatures to recognize and localize distant possibilities of interactions, and exhibits behaviors that satisfy its motivation principles, accordingly to this approach. In this paper, we propose a continual decision cycle between an agent and its environment to cope with the constraints that an artificial agent would meet in a physical environment.