CNN-based monocular decentralized SLAM on embedded FPGA. Yu, J., Gao, F., Cao, J., Yu, C., Zhang, Z., Huang, Z., Wang, Y., & Yang, H. Proceedings - 2020 IEEE 34th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2020, 2020. ![CNN-based monocular decentralized SLAM on embedded FPGA [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper doi abstract bibtex Decentralized visual simultaneous localization and mapping (DSLAM) can share locations and environmental information between robots, which is an essential task for many multi-robot applications. The visual odometry (VO) is a basic component to estimate the 6-DoF absolute pose for robot applications. Decentralized place recognition (DPR) is a fundamental element to produce candidate place matches for sharing information among different robots. The goal of this paper is to build a CNN-based real-time DSLAM system on embedded FPGA platforms. Because of the high precision requirement of VO, the existing quantization methods can not be directly applied. We improve the fixed-point fine-tune method for the CNN-based monocular VO, which enables VO can be deployed on the fixed-point FPGA accelerator. We also explore the influence of the DPR frequency on the DSLAM results, and find out a proper DPR frequency to balance the accuracy and speed. A cross-component pipeline scheduling method is proposed to improve DPR frequency and further improve the final accuracy of DSLAM under the same hardware resource constraints.
@article{
title = {CNN-based monocular decentralized SLAM on embedded FPGA},
type = {article},
year = {2020},
pages = {66-73},
id = {81481902-064b-3da0-91fa-111f1ddf8dd4},
created = {2022-08-19T09:21:33.034Z},
accessed = {2022-08-19},
file_attached = {true},
profile_id = {48fc0258-023d-3602-860e-824092d62c56},
group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},
last_modified = {2022-09-07T11:17:52.444Z},
read = {true},
starred = {true},
authored = {false},
confirmed = {true},
hidden = {false},
folder_uuids = {97693603-b330-4e3e-8cf5-d549e6474921},
private_publication = {false},
abstract = {Decentralized visual simultaneous localization and mapping (DSLAM) can share locations and environmental information between robots, which is an essential task for many multi-robot applications. The visual odometry (VO) is a basic component to estimate the 6-DoF absolute pose for robot applications. Decentralized place recognition (DPR) is a fundamental element to produce candidate place matches for sharing information among different robots. The goal of this paper is to build a CNN-based real-time DSLAM system on embedded FPGA platforms. Because of the high precision requirement of VO, the existing quantization methods can not be directly applied. We improve the fixed-point fine-tune method for the CNN-based monocular VO, which enables VO can be deployed on the fixed-point FPGA accelerator. We also explore the influence of the DPR frequency on the DSLAM results, and find out a proper DPR frequency to balance the accuracy and speed. A cross-component pipeline scheduling method is proposed to improve DPR frequency and further improve the final accuracy of DSLAM under the same hardware resource constraints.},
bibtype = {article},
author = {Yu, Jincheng and Gao, Feng and Cao, Jianfei and Yu, Chao and Zhang, Zhaoliang and Huang, Zhengfeng and Wang, Yu and Yang, Huazhong},
doi = {10.1109/IPDPSW50202.2020.00019},
journal = {Proceedings - 2020 IEEE 34th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2020}
}
Downloads: 0
{"_id":"Jz9gxLSiNGT9Qhw4n","bibbaseid":"yu-gao-cao-yu-zhang-huang-wang-yang-cnnbasedmonoculardecentralizedslamonembeddedfpga-2020","author_short":["Yu, J.","Gao, F.","Cao, J.","Yu, C.","Zhang, Z.","Huang, Z.","Wang, Y.","Yang, H."],"bibdata":{"title":"CNN-based monocular decentralized SLAM on embedded FPGA","type":"article","year":"2020","pages":"66-73","id":"81481902-064b-3da0-91fa-111f1ddf8dd4","created":"2022-08-19T09:21:33.034Z","accessed":"2022-08-19","file_attached":"true","profile_id":"48fc0258-023d-3602-860e-824092d62c56","group_id":"5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1","last_modified":"2022-09-07T11:17:52.444Z","read":"true","starred":"true","authored":false,"confirmed":"true","hidden":false,"folder_uuids":"97693603-b330-4e3e-8cf5-d549e6474921","private_publication":false,"abstract":"Decentralized visual simultaneous localization and mapping (DSLAM) can share locations and environmental information between robots, which is an essential task for many multi-robot applications. The visual odometry (VO) is a basic component to estimate the 6-DoF absolute pose for robot applications. Decentralized place recognition (DPR) is a fundamental element to produce candidate place matches for sharing information among different robots. The goal of this paper is to build a CNN-based real-time DSLAM system on embedded FPGA platforms. Because of the high precision requirement of VO, the existing quantization methods can not be directly applied. We improve the fixed-point fine-tune method for the CNN-based monocular VO, which enables VO can be deployed on the fixed-point FPGA accelerator. We also explore the influence of the DPR frequency on the DSLAM results, and find out a proper DPR frequency to balance the accuracy and speed. A cross-component pipeline scheduling method is proposed to improve DPR frequency and further improve the final accuracy of DSLAM under the same hardware resource constraints.","bibtype":"article","author":"Yu, Jincheng and Gao, Feng and Cao, Jianfei and Yu, Chao and Zhang, Zhaoliang and Huang, Zhengfeng and Wang, Yu and Yang, Huazhong","doi":"10.1109/IPDPSW50202.2020.00019","journal":"Proceedings - 2020 IEEE 34th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2020","bibtex":"@article{\n title = {CNN-based monocular decentralized SLAM on embedded FPGA},\n type = {article},\n year = {2020},\n pages = {66-73},\n id = {81481902-064b-3da0-91fa-111f1ddf8dd4},\n created = {2022-08-19T09:21:33.034Z},\n accessed = {2022-08-19},\n file_attached = {true},\n profile_id = {48fc0258-023d-3602-860e-824092d62c56},\n group_id = {5ec9cc91-a5d6-3de5-82f3-3ef3d98a89c1},\n last_modified = {2022-09-07T11:17:52.444Z},\n read = {true},\n starred = {true},\n authored = {false},\n confirmed = {true},\n hidden = {false},\n folder_uuids = {97693603-b330-4e3e-8cf5-d549e6474921},\n private_publication = {false},\n abstract = {Decentralized visual simultaneous localization and mapping (DSLAM) can share locations and environmental information between robots, which is an essential task for many multi-robot applications. The visual odometry (VO) is a basic component to estimate the 6-DoF absolute pose for robot applications. Decentralized place recognition (DPR) is a fundamental element to produce candidate place matches for sharing information among different robots. The goal of this paper is to build a CNN-based real-time DSLAM system on embedded FPGA platforms. Because of the high precision requirement of VO, the existing quantization methods can not be directly applied. We improve the fixed-point fine-tune method for the CNN-based monocular VO, which enables VO can be deployed on the fixed-point FPGA accelerator. We also explore the influence of the DPR frequency on the DSLAM results, and find out a proper DPR frequency to balance the accuracy and speed. A cross-component pipeline scheduling method is proposed to improve DPR frequency and further improve the final accuracy of DSLAM under the same hardware resource constraints.},\n bibtype = {article},\n author = {Yu, Jincheng and Gao, Feng and Cao, Jianfei and Yu, Chao and Zhang, Zhaoliang and Huang, Zhengfeng and Wang, Yu and Yang, Huazhong},\n doi = {10.1109/IPDPSW50202.2020.00019},\n journal = {Proceedings - 2020 IEEE 34th International Parallel and Distributed Processing Symposium Workshops, IPDPSW 2020}\n}","author_short":["Yu, J.","Gao, F.","Cao, J.","Yu, C.","Zhang, Z.","Huang, Z.","Wang, Y.","Yang, H."],"urls":{"Paper":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c/file/71a79fd1-d1d7-2cb5-73e4-972b40c566b2/full_text.pdf.pdf"},"biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","bibbaseid":"yu-gao-cao-yu-zhang-huang-wang-yang-cnnbasedmonoculardecentralizedslamonembeddedfpga-2020","role":"author","metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/service/mendeley/bfbbf840-4c42-3914-a463-19024f50b30c","dataSources":["2252seNhipfTmjEBQ"],"keywords":[],"search_terms":["cnn","based","monocular","decentralized","slam","embedded","fpga","yu","gao","cao","yu","zhang","huang","wang","yang"],"title":"CNN-based monocular decentralized SLAM on embedded FPGA","year":2020}