Affordance Equivalences in Robotics: A Formalism

Area of Science:

• Robotics
• Artificial Intelligence
• Cognitive Science

Background:

• Automatic knowledge grounding remains a challenge in cognitive robotics.
• Developmental robotics highlights the importance of environmental interaction for robots to learn action effects.
• Affordances, defined as actor-action-object-effect relationships, are key to this learning process.

Purpose of the Study:

• To propose a formalism for defining and identifying affordance equivalence in robots.
• To enable robots to acquire grounded knowledge by comparing affordances.
• To facilitate adaptive behavior by finding alternative action paths when actions or objects change.

Main Methods:

• Developing a formalism to compare elements of two affordances for equivalence.
• Identifying affordance equivalences to acquire grounded knowledge for robots.
• Validating identified affordance equivalences using recorded interaction data.

Main Results:

• The proposed formalism allows for the identification of equivalences between different affordances.
• This identification of equivalences contributes to the acquisition of grounded knowledge for robots.
• Experimental validation confirmed the coherence of interaction data with identified affordance equivalences.

Conclusions:

• Affordance equivalence provides a mechanism for robots to acquire grounded knowledge.
• This approach supports robot adaptability by enabling the discovery of alternative goal-reaching strategies.
• The study validates the effectiveness of using Bayesian Networks to verify affordance equivalences.