Short answers, pulled from the story.
Nouvelle AI is an approach to artificial intelligence pioneered in the 1980s by Rodney Brooks at MIT. Unlike classical symbolic AI, which programs robots with internal models of their world built from symbolic descriptions, Nouvelle AI lets robots rely continuously on their sensors rather than stored representations, allowing more complex behaviors to emerge from combinations of simpler ones.
Nouvelle AI was pioneered by Rodney Brooks, who was a researcher at MIT's artificial intelligence laboratory in the 1980s. Brooks developed the approach in reaction to the limitations of symbolic AI systems like Shakey and Freddy.
Brooks built several insectoid robots including Allen, named after Allen Newell, and Herbert, named after Herbert A. Simon. His team also built Genghis, a six-legged robot capable of walking over rough terrain, and Squirt, which hid in dark corners until it heard a noise. In the 1990s, Brooks and Lynn Andrea Stein built a humanoid robot called Cog.
Herbert was an insectoid robot built by Rodney Brooks that used infrared sensors and a laser system to collect 3D data over about 12 feet. It roamed the busy offices and workspaces of the MIT AI lab searching for and collecting empty soda cans, a seemingly goal-oriented behavior that emerged from 15 simple behavior units combining.
The frame problem describes the difficulty of using first-order logic to represent facts about a robot's world, specifically how to express that things in an environment do not change arbitrarily without requiring a large number of axioms. Nouvelle AI seeks to sidestep this problem by dispensing with symbolic language and instead letting complex behaviors emerge from simpler behavioral elements.
Cog was a humanoid robot built by Rodney Brooks and Lynn Andrea Stein in the 1990s. It featured an extensive collection of sensors, a face, and arms intended to let it learn common-sense knowledge through embodied experience. All development of the project had ceased by 2003.