Abstract
Currently, robotic grasping methods based on sparse partial point clouds have attained excellent grasping performance on various objects. However, they often generate wrong grasping candidates due to the lack of geometric information on the object. In this work, we propose a novel and robust sparse shape completion model (TransSC). This model has a transformer-based encoder to explore more point-wise features and a manifold-based decoder to exploit more object details using a segmented partial point cloud as input. Quantitative experiments verify the effectiveness of the proposed shape completion network and demonstrate that our network outperforms existing methods. Besides, TransSC is integrated into a grasp evaluation network to generate a set of grasp candidates. The simulation experiment shows that TransSC improves the grasping generation result compared to the existing shape completion baselines. Furthermore, our robotic experiment shows that with TransSC, the robot is more successful in grasping objects of unknown numbers randomly placed on a support surface.
Highlights
Robotic grasping evaluation is a challenging task due to incomplete geometric information from single-view visual sensor data [28]
Because the target object and stacking blocks are all placed on the table vertically and the horizontal length of each block is bigger than the maximum horizontal width of target object, the occlusion ratio is calculated through measuring the vertical height of stacking blocks (Hb) and target object (Ht )
We present a novel transformer-based sparse shape completion network (TransSC)
Summary
Robotic grasping evaluation is a challenging task due to incomplete geometric information from single-view visual sensor data [28]. With the development of deep learning techniques, data-driven grasp detection methods have shown great potential [4, 15, 22, 31] to solve this problem They generate lots of grasp candidates and. Object uncertainty still exists and extra sensor interference with the object will directly affect the final grasping result Another strategy is to use shape completion to infer the original object shape while traditional grasping-based shape completion methods use a high-resolution voxelized grid as object representation [17, 18, 27], causing a high memory cost and information loss due to the sparsity of the sensory input. To avoid extra sensor cost and obtain complete object information, 45 Page 2 of 14
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
