FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks

Kaijie Wei, Koki Honda, Hideharu Amano

研究成果: Conference contribution

抄録

We propose an autonomous vehicle controlled by FPGAs. In our design, considering embedded systems, we apply the binarized neural networks (BNNs) which can realize a satis-fying result in high speed and accuracy to recognize pedestrians and some obstacles on a given road. To detect the traffic light, a passive camera-based pipeline is applied. Furthermore, the implementation of road lane detection is based on color selection algorithm, Canny Edge Detection, and Hough Transformation. The proposed design is realized by two Xilinx boards: PYNQ-Z1 and Zynq-Xc7Z010. These two FPGA boards cooperate with each other through a shared network cable. In the proposed design, the resource used by Zynq-Xc7Z010 can be greatly reduced and the inference time on the FPGA has been thousands times faster than the software implementation.

元の言語English
ホスト出版物のタイトルProceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018
出版者Institute of Electrical and Electronics Engineers Inc.
ページ428-431
ページ数4
ISBN(電子版)9781728102139
DOI
出版物ステータスPublished - 2018 12 1
イベント17th International Conference on Field-Programmable Technology, FPT 2018 - Naha, Okinawa, Japan
継続期間: 2018 12 102018 12 14

出版物シリーズ

名前Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018

Conference

Conference17th International Conference on Field-Programmable Technology, FPT 2018
Japan
Naha, Okinawa
期間18/12/1018/12/14

Fingerprint

Field programmable gate arrays (FPGA)
Neural networks
Edge detection
Embedded systems
Telecommunication traffic
Cables
Pipelines
Cameras
Color

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Hardware and Architecture

これを引用

Wei, K., Honda, K., & Amano, H. (2018). FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks. : Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018 (pp. 428-431). [8742321] (Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/FPT.2018.00091

FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks. / Wei, Kaijie; Honda, Koki; Amano, Hideharu.

Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 428-431 8742321 (Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018).

研究成果: Conference contribution

Wei, K, Honda, K & Amano, H 2018, FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks. : Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018., 8742321, Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018, Institute of Electrical and Electronics Engineers Inc., pp. 428-431, 17th International Conference on Field-Programmable Technology, FPT 2018, Naha, Okinawa, Japan, 18/12/10. https://doi.org/10.1109/FPT.2018.00091
Wei K, Honda K, Amano H. FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks. : Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 428-431. 8742321. (Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018). https://doi.org/10.1109/FPT.2018.00091
Wei, Kaijie ; Honda, Koki ; Amano, Hideharu. / FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks. Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 428-431 (Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018).
@inproceedings{51f5cb84ff7742758a6d5fd9fe6753f4,
title = "FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks",
abstract = "We propose an autonomous vehicle controlled by FPGAs. In our design, considering embedded systems, we apply the binarized neural networks (BNNs) which can realize a satis-fying result in high speed and accuracy to recognize pedestrians and some obstacles on a given road. To detect the traffic light, a passive camera-based pipeline is applied. Furthermore, the implementation of road lane detection is based on color selection algorithm, Canny Edge Detection, and Hough Transformation. The proposed design is realized by two Xilinx boards: PYNQ-Z1 and Zynq-Xc7Z010. These two FPGA boards cooperate with each other through a shared network cable. In the proposed design, the resource used by Zynq-Xc7Z010 can be greatly reduced and the inference time on the FPGA has been thousands times faster than the software implementation.",
keywords = "Autonomous driving, Binarized Neural Networks, Canny Edge Detection, FPGA, Hough transformation, Image processing, Pynq, Zynq",
author = "Kaijie Wei and Koki Honda and Hideharu Amano",
year = "2018",
month = "12",
day = "1",
doi = "10.1109/FPT.2018.00091",
language = "English",
series = "Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "428--431",
booktitle = "Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018",

}

TY - GEN

T1 - FPGA Design for Autonomous Vehicle Driving Using Binarized Neural Networks

AU - Wei, Kaijie

AU - Honda, Koki

AU - Amano, Hideharu

PY - 2018/12/1

Y1 - 2018/12/1

N2 - We propose an autonomous vehicle controlled by FPGAs. In our design, considering embedded systems, we apply the binarized neural networks (BNNs) which can realize a satis-fying result in high speed and accuracy to recognize pedestrians and some obstacles on a given road. To detect the traffic light, a passive camera-based pipeline is applied. Furthermore, the implementation of road lane detection is based on color selection algorithm, Canny Edge Detection, and Hough Transformation. The proposed design is realized by two Xilinx boards: PYNQ-Z1 and Zynq-Xc7Z010. These two FPGA boards cooperate with each other through a shared network cable. In the proposed design, the resource used by Zynq-Xc7Z010 can be greatly reduced and the inference time on the FPGA has been thousands times faster than the software implementation.

AB - We propose an autonomous vehicle controlled by FPGAs. In our design, considering embedded systems, we apply the binarized neural networks (BNNs) which can realize a satis-fying result in high speed and accuracy to recognize pedestrians and some obstacles on a given road. To detect the traffic light, a passive camera-based pipeline is applied. Furthermore, the implementation of road lane detection is based on color selection algorithm, Canny Edge Detection, and Hough Transformation. The proposed design is realized by two Xilinx boards: PYNQ-Z1 and Zynq-Xc7Z010. These two FPGA boards cooperate with each other through a shared network cable. In the proposed design, the resource used by Zynq-Xc7Z010 can be greatly reduced and the inference time on the FPGA has been thousands times faster than the software implementation.

KW - Autonomous driving

KW - Binarized Neural Networks

KW - Canny Edge Detection

KW - FPGA

KW - Hough transformation

KW - Image processing

KW - Pynq

KW - Zynq

UR - http://www.scopus.com/inward/record.url?scp=85068317187&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068317187&partnerID=8YFLogxK

U2 - 10.1109/FPT.2018.00091

DO - 10.1109/FPT.2018.00091

M3 - Conference contribution

AN - SCOPUS:85068317187

T3 - Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018

SP - 428

EP - 431

BT - Proceedings - 2018 International Conference on Field-Programmable Technology, FPT 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -