Hand it to you: Our ability to grab, hold, and manipulate objects involves our dexterous hands, our sense of touch, and feedback from our eyes and muscles that allows us to maintain a controlled grip. Billard and Kragic review the progress made in robotics to emulate these functions. Systems have developed from simple, pinching grippers operating in a fully defined environment, to robots that can identify, select, and manipulate objects from a random collection. Further developments are emerging from advances in computer vision, computer processing capabilities, and tactile materials that give feedback to the robot. Science , this issue p. eaat8414
Grasping is the process of picking up an object by applying forces and torques at a set of contacts. Recent advances in deep learning methods have allowed rapid progress in robotic object grasping. In this systematic review, we surveyed the publications over the last decade, with a particular interest in grasping an object using all six degrees of freedom of the end-effector pose. Our review found four common methodologies for robotic grasping: sampling-based approaches, direct regression, reinforcement learning, and exemplar approaches In addition, we found two “supporting methods” around grasping that use deep learning to support the grasping process, shape approximation, and affordances. We have distilled the publications found in this systematic review (85 papers) into ten key takeaways we consider crucial for future robotic grasping and manipulation research.
We propose to leverage a real-world, human activity RGB dataset to teach a robot Task-Oriented Grasping (TOG). While RGB datasets contain example grasps of a large variety of objects for many different tasks, they often lack annotations for inferring a 6D robotic grasp pose. We propose a model that takes as input an RGB image and outputs a hand pose and configuration as well as an object pose and a shape. We follow the insight that jointly estimating hand and object poses increases accuracy compared to estimating these quantities independently of each other. Given the trained model, we process an RGB dataset to automatically obtain the data to train a TOG model. This model takes as input an object point cloud and outputs a suitable region for task-specific grasping. Our ablation study shows that training an object pose predictor with the hand pose information (and vice versa) is better than training without this information. Furthermore, our results on a real-world dataset show the applicability and competitiveness of our method over state-of-the-art. Experiments with a robot demonstrate that this data allows a robot to learn task-oriented grasping on novel objects.