Abstract
Reinforcement learning offers one of the most general frame-work to take traditional robotics towards true autonomy and versatility.
However, applying reinforcement learning to high dimensional movement
systems like humanoid robots remains an unsolved problem. In this pa-
per, we discuss different approaches of reinforcement learning in terms of
their applicability in humanoid robotics. Methods can be coarsely clas-
sified into three different categories, i.e., greedy methods, `vanilla‘ policy
gradient methods, and natural gradient methods. We discuss that greedy
methods are not likely to scale into the domain humanoid robotics as
they are problematic when used with function approximation. `Vanilla‘
policy gradient methods on the other hand have been successfully ap-
plied on real-world robots including at least one humanoid robot [3].
We demonstrate that these methods can be significantly improved us-
ing the natural policy gradient instead of the regular policy gradient. A
derivation of the natural policy gradient is provided, proving that the av-
erage policy gradient of Kakade [10] is indeed the true natural gradient.
A general algorithm for estimating the natural gradient, the Natural
Actor-Critic algorithm, is introduced. This algorithm converges to the
nearest local minimum of the cost function with respect to the Fisher in-
formation metric under suitable conditions. The algorithm outperforms
non-natural policy gradients by far in a cart-pole balancing evaluation,
and for learning nonlinear dynamic motor primitives for humanoid robot
control. It offers a promising route for the development of reinforcement
learning for truly high-dimensionally continuous state-action systems.
