Abstract:We propose an active learning-based grasp method for accurate industrial manipulation that combines the high accuracy of geometrically driven grasp methods and the generalization ability of data-driven grasp methods. Our grasp sequence consists of pregrasp stage and grasp stage which integrates the active perception and manipulation. In pregrasp stage, the manipulator actively moves and perceives the object. At each step, given the perception image, a motion is chosen so that the manipulator can adjust to a proper pose to grasp the object. We train a convolutional neural network to estimate the motion and combine the network with a closed-loop control so that the end effector can move to the pregrasp state. In grasp stage, the manipulator executes a fixed motion to complete the grasp task. The fixed motion can be acquired from the demonstration with nonexpert conveniently. Our proposed method does not require the prior knowledge of camera intrinsic parameters, hand-eye transformation, or manually designed feature of objects. Instead, the training data sets containing prior knowledge are collected through interactive perception. The method can be easily transferred to new tasks with a few human interventions and is able to complete high accuracy grasp task with a certain robustness to partial observation condition. In our circuit board grasping tests, we could achieve a grasp accuracy of 0.8 mm and 0.6°. Note to Practitioners—The research in this paper is motivated by the following practical problem. Manipulators on industrial lines can complete high accuracy tasks with hand-crafting features of objects. The perception is only used for object detection and localization. It is not flexible since the prior knowledge differs from tasks, which takes a long time to deploy in a new task. Besides, only well-trained experts are qualified to complete the deployment process. Our grasp method uses a convolutional network to estimate the motion for manipulator directly from images. The camera is mounted on the manipulator and can perceive the object actively. The training data set of the network is specific for different objects that can be automatically collected with a few human interventions. Our method simplifies the deployment process and can be applied in 3C industry (computers, communications, and consumer electronics) where the products upgrade frequently.

Grasp Type Understanding — Classification, Localization and Clustering

Active Learning-Based Grasp for Accurate Industrial Manipulation.

LiteGrasp: A Light Robotic Grasp Detection Via Semi-Supervised Knowledge Distillation

Attention Based Visual Analysis for Fast Grasp Planning with a Multi-Fingered Robotic Hand

Real-Time Robotic Multigrasp Detection Using Anchor-Free Fully Convolutional Grasp Detector

A Multi-Scale Grasp Detector Based on Fully Matching Model

Robotic Grasp Detection Using Structure Prior Attention and Multiscale Features

Two-stage Grasp Detection Method for Robotics Using Point Clouds and Deep Hierarchical Feature Learning Network

Robotic Grasp Detection Network Based on Improved Deformable Convolution and Spatial Feature Center Mechanism

Graspness Discovery in Clutters for Fast and Accurate Grasp Detection

Overview of Robotic Grasp Detection from 2D to 3D

Grasp Stability Prediction Using Tactile Information.

Object Classification and Grasp Planning Using Visual and Tactile Sensing

AnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domains

Dynamic Grasp Recognition Using Time Clustering, Gaussian Mixture Models and Hidden Markov Models

Glove-Net: Enhancing Grasp Classification with Multisensory Data and Deep Learning Approach

Predict Robot Grasp Outcomes Based On Multi-Modal Information

Grasp Prediction and Evaluation of Multi-Fingered Dexterous Hands Using Deep Learning.

Robotic grasping using visual and tactile sensing

Rethinking 6-Dof Grasp Detection: A Flexible Framework for High-Quality Grasping

A Real-Time Robotic Grasping Approach with Oriented Anchor Box