Learning robot in-hand manipulation with tactile features

Stable reinforcement learning with autoencoders for tactile and visual data