Henrik Brådland

• Brådland, Henrik; Choux, Martin Marie Hubert & Cenkeramaddi, Linga Reddy (2022). Point Cloud Instance Segmentation for Automatic Electric Vehicle Battery Disassembly. In Sanfilippo, Filippo; Granmo, Ole-Christoffer; Yildirim Yayilgan, Sule & Bajwa, Imran Sarwar (Ed.), Intelligent Technologies and Applications: 4th International Conference, INTAP 2021, Grimstad, Norway, October 11–13, 2021, Revised Selected Papers. Springer. ISSN 978-3-031-10524-1. p. 247–258. doi: 10.1007/978-3-031-10525-8_20. Show summary

  This paper describes a novel design based on recent 3D perception methods for capturing point clouds and segmenting instances of cabling found on electric vehicle battery packs. The use of cutting-edge perception algorithm architectures, such as graph-based and voxel-based convolution, in industrial autonomous lithium-ion battery pack disassembly is investigated. The proposed approach focuses on the challenge of getting a desirable representation of any battery pack using an industrial robot in conjunction with a high-end structured light camera, with “end-to-end” and “model-free” as design constraints. The proposed design employs self-captured datasets comprised of several battery packs that have been captured and labeled. Following that, the datasets are used to create a perception system. Based on the results, graph-based deep-learning algorithms have been shown to be capable of being scaled up to 50, 000 inputs while still exhibiting strong performance in terms of accuracy and processing time. The results show that an instance segmenting system can be implemented in less than two seconds. Using off-the-shelf hardware, we demonstrate that a 3D perception system is industrially viable and competitive as compared to a 2D perception system (The different algorithms studied in this article are implemented in Python and can be obtained through the following link: https://github.com/HenrikBradland-Nor/intap21).

View all works in Cristin