The ability to grasp objects is an essential skill that enables many robotic manipulation tasks. Recent works have studied point cloud-based methods for object grasping by starting from simulated datasets and have shown promising performance in real-world scenarios. Nevertheless, many of them still rely on ad-hoc geometric heuristics to generate grasp candidates, which fail to generalize to objects with significantly different shapes with respect to those observed during training. Several approaches exploit complex multi-stage learning strategies and local neighborhood feature extraction while ignoring semantic global information. Furthermore, they are inefficient in terms of number of training samples and time required for inference. In this letter, we propose an end-to-end learning solution to generate 6-DOF parallel-jaw grasps starting from the 3D partial view of the object. Our Learning to Grasp (L2G) method gathers information from the input point cloud through a new procedure that combines a differentiable sampling strategy to identify the visible contact points, with a feature encoder that leverages local and global cues. Overall, L2G is guided by a multi-task objective that generates a diverse set of grasps by optimizing contact point sampling, grasp regression, and grasp classification. With a thorough experimental analysis, we show the effectiveness of L2G as well as its robustness and generalization abilities.
The Industrial Assembly Challenge at the World Robot Summit was held in 2018 to showcase the state-of-the-art of autonomous manufacturing systems. The challenge included various tasks, such as bin picking, kitting, and assembly of standard industrial parts into 2D and 3D assemblies. Some of the tasks were only revealed at the competition itself, representing the challenge of ‘level 5’ automation, i.e. programming and setting up an autonomous assembly system in less than one day. We conducted a survey among the teams that participated in the challenge and investigated aspects such as team composition, development costs, system setups as well as the teams' strategies and approaches. An analysis of the survey results reveals that the competitors have been in two camps: those constructing conventional robotic work cells with off-the-shelf tools, and teams who mostly relied on custom-made end effectors and novel software approaches in combination with collaborative robots. While both camps performed reasonably well, the winning team chose a middle ground in between, combining the efficiency of established play-back programming with the autonomy gained by CAD-based object detection and force control for assembly operations.
This paper describes our concept for control of robots and machine tools through a private factory cloud which is dynamically extended by additional public resources. After discussing the opportunities and drawbacks of the virtualization of robot controllers (RC) and software programmable logic controllers (soft-PLC), we present first experiences with the setup of private and public infrastructures. Finally, we introduce our experimental design for investigating the requirements of machine control algorithms regarding availability, dependability and realtime capability in local and public configurations.
In this paper we contribute to the research on facilitating industrial robot programming by presenting a concept for intuitive drag and drop like programming of pick and place tasks with Augmented Reality (AR). We propose a service-oriented architecture to achieve easy exchangeability of components and scalability with respect to AR devices and robot workplaces. Our implementation uses a HoloLens and a UR5 robot, which are integrated into a framework of RESTful web services. The user can drag recognized objects and drop them at a desired position to initiate a pick and place task. Although the positioning accuracy is unsatisfactory yet, our implemented prototype achieves most of the desired advantages to proof the concept.
Deep Learning-based approaches for 3d object detection and 6d pose estimation typically require large amounts of labeled training data. Labeling image data is expensive and particularly the 6d pose information is difficult to obtain, as it requires a complex setup during image acquisition. Training with synthetic data is therefore very attractive. Large amounts of synthetic, labeled data can be generated, but it is not yet fully understood how certain aspects of data generation affect the detection and pose estimation performance. Our work therefore focuses on creating synthetic training data and investigating the effects on detection performance. We present two methods for data generation: rendering object views and pasting them on random background images, and simulating realistic scenes. The former is computationally simpler and achieved better results, but the detection performance is still very sensitive to small changes, e.g. the type of background images.
23 November 2019 in Lecture Notes in Computer Science