Rodney Brooks introduced the Nouvelle AI approach in the 1980s while working at the MIT artificial intelligence laboratory. He sought to solve persistent issues with earlier systems like Shakey and Freddy that relied on internal symbolic models of their environments.
The core idea was simple: use sensors to process information only when needed instead of maintaining a static internal database. Intelligence should emerge from direct interaction with the physical world rather than abstract symbol manipulation.
Development continued until 2003 when all work on the project ceased. The experiment demonstrated both the potential and the immense difficulty of scaling reactive behaviors into full human cognition.
The testing ground for these devices was the busy office space of the MIT AI lab where they searched for empty soda cans. These machines contained no internal models of their surroundings and discarded most sensory information within two seconds.
Traditional logic systems struggled with what researchers called the frame problem when trying to describe how objects behave over time. First-order logic required writing thousands of axioms just to state that things do not change arbitrarily, creating an impossible burden for any robot attempting to navigate a dynamic environment.