Learning robot in-hand manipulation with tactile features

TLDR

Dexterous manipulation enables repositioning of objects in a robot's hand, but for unknown objects exact models are unavailable, so compliance and tactile feedback are used to adapt, though compliant hands and sensors add modeling complexity. The study proposes acquiring in‑hand manipulation skills via reinforcement learning that does not rely on analytic dynamics or kinematics models. They use reinforcement learning to acquire tactile manipulation skills with a passively compliant hand, avoiding the need for analytic dynamics or kinematics models. The approach successfully acquires a tactile manipulation skill with a passively compliant hand and generalizes to novel objects.

Abstract

Dexterous manipulation enables repositioning of objects and tools within a robot's hand. When applying dexterous manipulation to unknown objects, exact object models are not available. Instead of relying on models, compliance and tactile feedback can be exploited to adapt to unknown objects. However, compliant hands and tactile sensors add complexity and are themselves difficult to model. Hence, we propose acquiring in-hand manipulation skills through reinforcement learning, which does not require analytic dynamics or kinematics models. In this paper, we show that this approach successfully acquires a tactile manipulation skill using a passively compliant hand. Additionally, we show that the learned tactile skill generalizes to novel objects.