High precision grasp pose detection in dense clutter

This paper considers the problem of grasp pose detection in point clouds. We follow a general algorithmic structure that first generates a large set of 6-DOF grasp candidates and then classifies each of them as a good or a bad grasp. Our focus in this paper is on improving the second step by using depth sensor scans from large online datasets to train a convolutional neural network. We propose two new representations of grasp candidates, and we quantify the effect of using prior knowledge of two forms: instance or category knowledge of the object to be grasped, and pretraining the network on simulated depth data obtained from idealized CAD models. Our analysis shows that a more informative grasp candidate representation as well as pretraining and prior knowledge significantly improve grasp detection. We evaluate our approach on a Baxter Research Robot and demonstrate an average grasp success rate of 93% in dense clutter. This is a 20% improvement compared to our prior work.

Authors

• Marcus Gualtieri
• Andreas ten Pas
• Kate Saenko
• Robert Platt 0001

Keywords

• Three-dimensional displays
• Solid modeling
• Geometry
• Robot sensing systems
• Grasping
• Grippers
• Grasp Pose Detection
• Convolutional Neural Network
• Simulated Data
• Point Cloud
• Depth Camera
• CAD Model
• Online Dataset
• Simulated Depth
• Training Data
• Classification Accuracy
• Channel Information
• Depth Images
• Depth Information
• Robotic Arm
• Object Surface
• Registration Error
• Round Of Experiments
• Object Dataset
• Robot Operating System
• Surface Normals
• Robotic Hand
• Point Cloud Registration
• RGB-D Images
• Kinect Sensor
• RGB Information
• Robot Experiments