Print Email Facebook Twitter Using Retinanet to determine local graspability for a suction actuator Title Using Retinanet to determine local graspability for a suction actuator Author van Tilburg, Floris (TU Delft Mechanical, Maritime and Materials Engineering) Contributor Hernández, Carlos (mentor) Wisse, M. (mentor) Kooij, J.F.P. (graduation committee) Imre, M. (graduation committee) Degree granting institution Delft University of Technology Date 2021-08-26 Abstract Suction based robotic actuators have potential for the bin-picking industry, but are currently not usable due the needed speed, accuracy and ability to handle novel and adversarial objects. An evaluation of the state of the art grasp pipeline developed by Mahler et al. [1] for detecting grasps on novel objects lead us to split the problem of robotic grasp generation into a global and a local component. The state of the art solution had the ability to fill the role of global evaluator leaving a local evaluator to be developed. This local grasp evaluator was obtained by training a neural network on a suction based grasp dataset, which was created using a newly developed annotation tool. The proposed grasp pipeline obtained by combining these two showed a 95.27% pick success rate for a random setup and a 90.28% success rate for solely novel objects. Subject suction gripperbin-pickingunknown objectsrobot visiondeep learning To reference this document use: http://resolver.tudelft.nl/uuid:d9e2c738-87b0-4bd9-b56a-7dfe2008f162 Part of collection Student theses Document type master thesis Rights © 2021 Floris van Tilburg Files PDF Msc_thesis_report_FAR_van ... 166169.pdf 58.27 MB Close viewer /islandora/object/uuid:d9e2c738-87b0-4bd9-b56a-7dfe2008f162/datastream/OBJ/view