Journal of IKEEE (전기전자학회논문지)

Institute of Korean Electrical and Electronics Engineers (한국전기전자학회)
권호 표지
DOI QR코드

DOI QR Code

Design and Implementation of Radar Signal Processing System for Vehicle Door Collision Prevention

차량 도어 충돌 방지용 레이다 신호처리 시스템 설계 및 구현

  • Jeongwoo Han (School of Electronics and Information Engineering, Korea Aerospace University) ;
  • Minsang Kim (School of Electronics and Information Engineering, Korea Aerospace University) ;
  • Daehong Kim (School of Electronics and Information Engineering, Korea Aerospace University) ;
  • Yunho Jung (School of Electronics and Information Engineering, Korea Aerospace University)
  • 한정우 ;
  • 김민상 ;
  • 김대홍 ;
  • 정윤호
  • Received : 2024.09.11
  • Accepted : 2024.09.24
  • Published : 2024.09.30
Download PDF
⟨ Previous Next ⟩

Abstract

This paper presents the design and implementation results of a Raspberry-Pi-based embedded system with an FPGA accelerator that can detect and classify objects using an FMCW radar sensor for preventing door collision accidents in vehicles. The proposed system performs a radar sensor signal processing and a deep learning processing that classifies objects into bicycles, automobiles, and pedestrians. Since the CNN algorithm requires substantial computation and memory, it is not suitable for embedded systems. To address this, we implemented a lightweight deep learning model, BNN, optimized for embedded systems on an FPGA, and verified the results achieving a classification accuracy of 90.33% and an execution time of 20ms.

본 논문에서는 차량의 개문사고를 예방하기 위한 목적으로 FMCW 레이다 센서를 활용하여 물체를 감지하고 분류 가능한 시스템 설계 및 구현 결과가 제시된다. 제안된 시스템은 Raspberry-Pi 기반 임베디드시스템과 FPGA 가속기에 기반하여 구현되었으며, 해당 시스템은 레이다 센서 신호처리 과정과 물체를 자전거, 자동차, 사람으로 분류하는 딥러닝 과정을 수행한다. CNN 알고리즘은 연산량과 메모리 사용량이 크기 때문에 임베디드시스템에 적합하지 않다. 이를 해결하기 위해 임베디드시스템에 적합한 경량화된 딥러닝 모델인 BNN을 FPGA 상에 구현한 뒤 결과를 검증하였고, 90.33%의 분류 정확도와 20ms의 수행시간을 확인하였다.

Keywords

Acknowledgement

This work was supported by the Technology Innovation Program (No. 00144290, 00433615), funded by the Ministry of Trade, Industry and Energy (MOTIE, Korea) and CAD tools were supported by IDEC.

References

  1. Z. -Y. Zhung, K. -C. Chen, Y. -H. Yu and N. Kwok, "Chip-based Anti-collision System for Car Door Opening," 2019 4th International Conference on Intelligent Transportation Engineering (ICITE), pp.322-326, 2019. DOI: 10.1109/ICITE.2019.8880149 
  2. Q. Shen, "Design of Backward Collision Warning and Avoidance System when On-street Parking Using LiDAR," 2021 2nd International Conference on Computing and Data Science (CDS), pp.239-242, 2021. DOI: 10.1109/CDS52072.2021.00048 
  3. Courbariaux, Matthieu, et al. "Binarized neural networks: Training deep neural networks with weights and activations constrained to+ 1 or-1." arXiv preprint arXiv:1602.02830, 2016. DOI: 10.48550/arXiv.1602.02830 
  4. G. -R. Liu, M. -Z. Zhou, L. -L. Wang and H. Wang, "A Radar-Based Door Open Warning Technology for Vehicle Active Safety," 2016 International Conference on Information System and Artificial Intelligence (ISAI), pp.479-484, 2016. DOI: 10.1109/ISAI.2016.0108 
  5. Yunsung Sim, Seungjun Song, Seonyoung Jang and Yunho Jung, "Design and Implementation of Human and Object Classification System Using FMCW Radar Sensor,"Journal of IKEEE, 26.3, pp. 364-372, 2022. DOI: 10.7471/ikeee.2022.26.3.364 
  6. E. Hyun and J. -H. Lee, "Multi-target tracking scheme using a track management table for automotive radar systems," 2016 17th International Radar Symposium (IRS), pp.1-5, 2016. DOI:10.1109/IRS.2016.7497283 
  7. Analog Devices, "EVAL-DEMORAD", https://www.analog.com/en/resources/evaluation-hardware-and-software/evaluation-boards-kits/evaldemorad.html#eb-overview 
  8. Raspberry Pi, "raspberry-pi-4-model-b", https://www.raspberrypi.com/products/raspberry-pi-4-model-b/ 
  9. terasic Technologies, "DE10 - Lite Board", https://www.terasic.com.tw/cgi-bin/page/archive.pl?Language=English&No=1021