FMCW 레이다 센서 기반 사람과사물 분류 시스템 설계 및 구현 = Design and Implementation of Human and Object Classification System Using FMCW Radar Sensor

다국어 초록 (Multilingual Abstract)

This paper proposes the design and implementation results for human and object classification systems utilizingfrequency modulated continuous wave (FMCW) radar sensor. Such a system requires the process of radar sensorsignal processing for multi-target detection and the process of deep learning for the classification of human andobject. Since deep learning requires such a great amount of computation and data processing, the lightweightprocess is utmost essential. Therefore, binary neural network (BNN) structure was adopted, operating convolutionneural network (CNN) computation in a binary condition. In addition, for the real-time operation, a hardwareaccelerator was implemented and verified via FPGA platform. Based on performance evaluation and verifiedresults, it is confirmed that the accuracy for multi-target classification of 90.5%, reduced memory usage by 96.87%compared to CNN and the run time of 5ms are achieved.

국문 초록 (Abstract)

본 논문에서는 FMCW(frequency modulated continuous wave) 레이다 센서를 활용한 사람과 사물을 분류하는 시스템 설계및 구현 결과를 제시한다. 해당 시스템은 다중 객체 탐지를 위한 레이다 센서 신호처리 과정과 객체를 사람 및 사물로 분류하는 딥러닝 과정을 수행한다. 딥러닝의 경우 높은 연산량과 많은 양의 메모리를 요구하기 때문에 경량화가 필수적이다. 따라서 CNN(convolution neural network) 연산을 이진화하여 동작하는 BNN (binary neural network) 구조를 적용하였으며, 실시간 동작을위해 하드웨어 가속기를 설계하고 FPGA 보드 상에서 구현 및 검증하였다. 성능 평가 및 검증 결과 90.5%의 다중 객체 구분 정확도,CNN 대비 96.87% 감소된 메모리 구현이 가능하며, 총 수행 시간은 5ms로 실시간 동작이 가능함을 확인하였다.

참고문헌 (Reference)

1 장대호 ; 김현 ; 정윤호, "실내 보안 응용을 위한 사람 감지 레이다 시스템의설계 및 구현" 한국전기전자학회 24 (24): 783-790, 2020

2 공영주 ; 우선걸 ; 박성호 ; 신승용 ; 장윤희 ; 양은정, "레이다 시스템 실시간 적용을 위한 OS CFAR 연산 시간 단축 방안" 한국전자파학회 29 (29): 791-798, 2018

3 Raspberry Pi, "raspberry-pi-4-model-b"

4 Hermann Rohling, "Radar CFAR thresholding in clutter and multiple target situations" Institute of Electrical and Electronics Engineers (IEEE) AES-19 (AES-19): 608-621, 1983

5 Md Alamgir Hossain, "Performance Assessment of Background Subtraction Algorithm" 1049-1051, 2019

6 H. Lee, "Path loss compensation method for multiple target vital sign detection with 24-GHz FMCW radar" 100-101, 2018

7 H. Rohling, "Ordered statistic cfar technique-an overview" 631 (631): 631-638, 2011

8 Abdul Jalil, "Object motion detection in home security system using the binary-image comparison method based on robot operating system 2 and Raspberry Pi" Universitas Muslim Indonesia 13 (13): 1-8, 2021

9 H. Rohling, "New radar waveform based on a chirp sequence" 1 (1): 1-4, 2014

10 E. Hyun, "Multi-target tracking scheme using a track management table for automotive radar systems" 1-5, 2016

11 I. G. M. N. Desnanjaya, "Home security monitoring system with IoT-based Raspberry Pi" Institute of Advanced Engineering and Science 22 (22): 1295-1302, 2021

12 Matthias Kronauge, "Fast Two-Dimensional CFAR Procedure" Institute of Electrical and Electronics Engineers (IEEE) 49 (49): 1817-1823, 2013

13 Infineon Technologies, "DEMO DISTANCE2GO"

14 terasic Technologies, "DE1 – SoC Board"

15 Nagesh Singh Chauhan, "DBSCAN Clustering Algorithm in Machine Learning"

16 남정희 ; 강채영 ; 국정연 ; 정윤호, "CW 레이다 기반 사람 행동 인식 시스템 설계 및 구현" 한국항행학회 25 (25): 415-425, 2021

17 Guochao Wang, "Application of Linear-Frequency-Modulated Continuous-Wave (LFMCW) Radars for Tracking of Vital Signs" Institute of Electrical and Electronics Engineers (IEEE) 62 (62): 1387-1399, 2014

18 Martin Ester, "A density-based algorithm for discovering clusters in large spatial databases with noise" 226-231, 1996

19 Qiangwen Zheng, "A Target Detection Scheme With Decreased Complexity and Enhanced Performance for Range-Doppler FMCW Radar" Institute of Electrical and Electronics Engineers (IEEE) 70 : 1-13, 2021

20 Hieu Vo Quang, "A Deep Learning Accelerator for a Typical Binary Neural Network Layer" 597-599, 2021