Automatic Development from Pixel-level Representation to Action-level Representation in Robot Navigation.
Jefferson Provost · 2007
Many important real-world robotic tasks have high diam-eter, that is, their solution requires a large number of prim-itive actions by the robot. For example, they may require navigating to distant locations using primitive motor control commands. In addition, modern robots are endowed with rich, high-dimensional sensory systems, providing measure-ments of a continuous environment. Reinforcement learning (RL) has shown promise as a method for automatic learning of robot behavior, but current methods work best on low-diameter, low-dimensional tasks. Because of this problem, the success of RL on real-world tasks still depends on hu-man analysis of the robot, environment, and task to provide a useful sensorimotor representation to the learning agent. A new method, Self-Organizing Distinctive-state Ab-straction (SODA) Provost, Kuipers, & Miikkulainen (2006);