Toward Sim-to-Real Directional Semantic Grasping

TL;DR

This paper presents a deep reinforcement learning approach for directional semantic grasping, using simulated data and domain randomization to enable real-world application of robot grasping from specific directions.

Contribution

It introduces an end-to-end system that maps RGB images to robot commands for directional grasping, bridging the sim-to-real gap with domain randomization.

Findings

Successful simulation and real-world grasping demonstrations

Effective use of DDQN and CEM for control

Identification of challenges for future research

Abstract

We address the problem of directional semantic grasping, that is, grasping a specific object from a specific direction. We approach the problem using deep reinforcement learning via a double deep Q-network (DDQN) that learns to map downsampled RGB input images from a wrist-mounted camera to Q-values, which are then translated into Cartesian robot control commands via the cross-entropy method (CEM). The network is learned entirely on simulated data generated by a custom robot simulator that models both physical reality (contacts) and perceptual quality (high-quality rendering). The reality gap is bridged using domain randomization. The system is an example of end-to-end (mapping input monocular RGB images to output Cartesian motor commands) grasping of objects from multiple pre-defined object-centric orientations, such as from the side or top. We show promising results in both simulation and the real world, along with some challenges faced and the need for future research in this area.