Grounding Artificial General Intelligence with Robotics: The PetitCat Project

The symbol grounding problem is a well-known philosophical problem. In the engineering of Artificial General Intelligence (AGI) systems, grounding becomes essential for interacting with a real-world environment. Grounding an AGI system or Brain-Inspired Cognitive Architecture (BICA) via mobile robotics is described. The PetitCat project includes an open-source software interface between the Python code of an AGI project or BICA project and a mobile robotics device built with open-source hardware. The interface allows real-time communication between the Python code on a desktop or server via wireless network to the robot’s lower-level C/C++ compiled code. The PetitCat project is targeted at studying symbol grounding and autonomous bottom-up interactive learning in the framework of theories of enaction, active inference, predictive coding, and developmental and constructivist learning.