In Cognitive Science, an affordance is a resource or support that the environment offers an agent for action, which the agent can directly perceive and employ. Only rarely has this concept been used in Robotics and AI, although it offers an original perspective on coupling perception, action and reasoning, differing from standard hybrid robot control architectures.
Perceiving affordances in the environment means perception as filtered through the individual capabilities for physical action and through the current goals or intentions, thereby coupling perception and action deep down in the control and providing an action-oriented interpretation of percepts in real time. Moreover, affordances provide on a high granularity level a basis for human-robot interaction and for learning or adapting context-dependent, goal-directed action.
The MACS project aims at developing affordance-based control as a new architecture paradigm for embedded cognitive systems. To this end, standard control architectures have to be completely redesigned to employ affordances effectively and efficiently in all modules of an affordance-based system. Accordingly, major lines of work in MACS address in the context of affordance-based control the topics of architecture, perception, representation, and learning. An autonomous mobile robot serves as an integrated demonstrator, which uses vision and 3D geometry sensing (3D laser scanner) as its main sensors and has locomotion and a simple manipulation device (2 DOF arm with a magnetic gripper) as bases for physical action.
In the course of the project, the physical robot demonstrator as well as its simulation will serve as an environment for integrating incremental versions of the software modules under development, providing a permanently available workbench for testing and evaluating versions of the architecture and of its modules.
Funding SchemeSTREP - Specific Targeted Research Project
581 83 Linkoeping