118
click to vote
ICRA
2005
IEEE
15 years 8 months ago
2005
IEEE
— This paper presents a new reinforcement learning algorithm for accelerating acquisition of new skills by real mobile robots, without requiring simulation. It speeds up Q-learni...