Jihoon Kwon

Jump to: navigation, search
Jihoon.jpg Jihoon Kwon (권지훈)

Ph. D. candidate
Program in Intelligent Systems
Department of Transdisciplinary Studies
Graduate School of Convergence Science and Technology
Seoul National University
e-mail: jihoonkwon@snu.ac.kr

Jihoon Kwon received his BS degree in Electrical and Computer Engineering at Chungnam National University, Korea, in 2004. He received an M.S. degree in School of Mechatronics at Gwangju Institute of Science and Technology (GIST), Gwangju, Korea, in 2007. He is currently Ph. D. Student in Graduate School of Convergence Science and Technology at Seoul National University, Korea. He also has been working in the R&D center of Hanwha systems company (Samsung Thales) since 2007. His research interests are in the area of cognitive radar, advanced tracking filter, machine learning & pattern recognition, radar signal processing and multi-sensor fusion.

My future works: When we solve radar pattern recognition problems, the most bothersome problem is that the amount of data is insufficient. So, we should have a lot of interest in the generative models that can make virtual dataset effectively. We expect that Variational Autoencoder (VAE) and Generative Adversarial Networks (GAN) which are recently introduced in the field of Computer Vision are going to be good alternatives to overcome this restriction. I think that the studies related to implementation of generative models for radar machine learning are going to be very interesting, meaningful and challenging researches.

Also, I have a lot of interests in convergence of the advanced tracking filter and the deep learning algorithm. In the previous most of researches, the tracking filter has been considered as the Baysian estimation problem. However, in the future, since we have to consider extremely low SNR (Signal to Noise Ratio) targets, I think that the conventional approaches have limitations to detect and track a target. So, we have to suggest new approaches that can detect targets with extemely low SNR. I think that the radar machine learning technology related to re-using dataset or database model captured by a radar can effectively improve SNR of targets. Also I expect that these ideas will give many inspiration to young engineer like me.


  • Present:
    Ph.D. Student in Graduate School of Convergence Science and Technology, Seoul National University, Korea.
  • Feb. 2007:
    M.S. in School of Mechatronics, Gwangju Institute of Science and Technology (GIST), Korea.
    Research Topics: W-band Radar & Radiometer Sensor System for Detection of Hidden Objects, V-band RF Front-end for High-speed Communications
  • Aug. 2004:
    B.S. in Electrical and Computer Engineering, Chungnam National University, Korea.


  • Jan. 2017 ~ Present: Senior Engineer, Multi-Function Radar Team, Hanwha Systems.
  • Jan. 2013 ~ 2016: Senior Engineer, Radar-EW R&D Center, Hanwha Systems.
  • Jan. 2007 ~ Dec.2012: Engineer, Radar System Group, Samsung Thales.


International Journals
    [1] Y. Kim, S. Ha and J. Kwon, “Human detection using Doppler radar based on physical characteristics of targets,” IEEE Geoscience and Remote Sensing Letters, vol.12, pp. 289 – 293, Feb. 2015.
    [2] J. Bryan, J. Kwon, N. Lee and Y. Kim, “Application of UWB radar for classification of human activities,” IET Radar, Sonar and Navigation, vol.6, pp. 172–179, Mar. 2012.
    [3] C. Chae, J. Kwon, and Y. Kim, "A Study of Compensation for Temporal and Spatial Physical Temperature Variation in Total Power Radiometers ," Sensors Journal, IEEE, vol.12, pp: 2306 - 2312, 2012
    [4] C. Chae, J. Kwon, and Y. Kim, "Ka-band total power radiometer and temperature control system for calibration of spatial temperature distribution," Sensors Journal, IEEE, vol.10, pp: 1555 - 1556, 2010

International Conferences
    [1] Jihoon Kwon and Nojun Kwak, "Radar Detection Algorithm Based on Deep Learning for Recgnizing Micro-Doppler Signals by Human Walking and Background Noise", International Symposium on Antenna and Propagation (ISAP 2018), Pusan, Republic of Korea. (Invited, Preparation)
    [2] Jihoon Kwon, Nojun Kwak, Eunjeoung Yang and Kwonsung Kim, "Ensemble Kalman filter for Track-before-detect algorithm of pulse-Doppler radar", The 8th Microwave and Radar Week, Poznan, Poland, May 2018. (Oral presentation, Accepted)
    [3] Jihoon Kwon, Nojun Kwak, Eunjeoung Yang and Kwonsung Kim, "Weight Assignment Data Association Algorithm of Ensemble Kalman Filter for Pulse-Doppler Radar", International Conference on Radar, Brisbane, Australia, August 2018. (Submitted)
    [4] Jihoon Kwon and Nojun Kwak "Human Detection by Deep Neural Networks Recognizing Micro-Doppler Signals of Radar", European Microwave Week 2018, Madrid, Spain, September 2018. (Oral presentation, Accepted)
    [5] Jihoon Kwon, Nojun Kwak, Dongwon Yang and Yeosun Yoon, "Development of Ground Penetrating Radar based on 48 Channel Impulse Radar Array", European Microwave Week 2017, Nuremberg, Germany, October 2017. (Oral presentation)
    [6] Jihoon Kwon, Nojun Kwak, Yeosun Yoon and Dongwon Yang, "Development of GPR Using 16 Channel Impulse UWB Radar Array", 2017 IEEE International AP-S symposium, San Diego, CA, July 2017. (Oral presentation)
    [7] Jihoon Kwon and Nojun Kwak, "Human detection by neural networks using a low-cost short-range Doppler radar sensor", 2017 IEEE Radar Conference, Seattle, WA, May 2017. (Oral presentation)

Domestic Journals
    [1] 권지훈, 곽노준 "사람 걸음 탐지 및 배경잡음 분류 처리를 위한 도플러 레이다용 딥뉴럴네트워크", 한국전자파학회논문지, 2018. (Submitted)
    [2] Jeehyun Lee, Jihoon Kwon, Jin-Ho Bae, and Chong Hyun Lee,* "Classification Algorithms for Human and Dog Movement Based on Micro-Doppler Signals", IEIE Transactions on Smart Processing and Computing, vol. 6, no. 1, pp. 10∼17, February 2017
    [3] 권지훈, 곽노준, 하성재, 한승훈, 윤여선, 양동원, "지뢰탐지용 48채널 배열 UWB 임펄스 레이더 방식 지면투과레이더시스템 개발", 대한전자공학회, Vol. 53, No. 12, pp. 1765∼1774, Dec. 2016.
    [4] 권지훈, 강성철, 곽노준, "TBD 처리를 위한 레이더용 파티클필터 기법 연구", 한국전자파학회논문지, Vol. 27, No. 3, pp. 317∼325, Mar. 2016.
    [5] 이재일, 이종현, 배진호, 권지훈, ”배경잡음에 적응하는 진동센서 기반 목표물 탐지 알고리즘”, 전자공학회논문지, vol.50, No. 7, 2013

Domestic Conference Papers
    [1] 권지훈, 곽노준, "도플러 레이더 패턴 인식을 통한 인간 보행 감지 및 배경잡음 제거 기법", 한국자동차공학회 춘계종합학술대회 제출, 2016년
    [2] 권지훈, 김동현, 한승훈, “초광대역 임펄스 신호를 사용한 지면투과레이더 매설표적 영상획득 연구”, 대한전자공학회 하계종합학술대회, vol. 2014, No.6, 2014
    [3] 권지훈, 김동현, 한승훈, “투과레이더용 초광대역 임펄스 방식 송수신기 개발”, 한국군사과학기술학회, 종합학술대회, 2014
    [4] 권지훈, 한승훈, 윤여선, “초광대역 투과레이더, 자성 및 중성자 센서 복합기반 지뢰탐지장치 개발 개요”, 한국군사과학기술학회, 2013
    [5] 권지훈 외 ”실외형 마이크로-도플러레이더 센서 하드웨어 및 탐지처리기법 개발”, 한국군사과학기술학회, 2013 외 4편

    [1] 매설물 탐지 레이더 및 탐지 방법, Korean patent, Register No. 1015518240000, 2015
    [2] 진동센서 기반의 능동형 문턱값을 이용한 표적 탐지 시스템 및 이를 이용한 표적 탐지 방법, Korean patent, Register No. 1013020600000, 2013
    [3] 테이퍼드 슬롯 안테나 장치 및 이를 구비한 레이더, Korean patent, Register No. 1012122190000, 2012

Under Preparation

I am currently focusing on writing SCI journals related to the radar deep learning and the advanced tracking filter. I expect to get good results at the end of 2018 or the beginning of 2019. In 2019-2020, I will focus on the generative models for radar machine learning algorithms.

International Journals

    [1] Jihoon Kwon and Nojun Kwak, "A Novel Radar Detection Method Based on Deep Learning for Pulsed-Doppler Radar in Heavy Clutter Environments", - , 2018~2019
    This topic will be studied as my doctoral dissertation. This journel will explain an effective detection method that uses machine learning algorithms instead of CFAR (Constant False Alarm Rate). I hope to suggest a novel detection method that can replace CFAR and this suggested algorithm is going to be effectively applied in a heavy clutter environment.
    [2] Jihoon Kwon and Nojun Kwak, "A Novel Data Association Algorithm for Ensemble Kalman Filter", - , 2018
    In the field of radar, the particle filter have been studied as an effective filter to improve non-linearity. On the other hand, the ensemble Kalman filter is very similar to the particle filter. But, in the field of radar, it has been studied less importantly than the particle filter. The purpose of this study is to investigate the effectiveness of the ensemble Kalman filter and to propose an efficient data association algorithm to improve the computation complexity.
    [3] Jihoon Kwon and Nojun Kwak, "Application of Radar Deep Learning for Recognizing Human Activities from Noisy Micro-Doppler Signals", - , 2018
    For several years, I have researched the method for recognizing the Micro-Doppler signals, and I showed that the deep neural networks can be applied as an effective approach to classfy human walking motions from background noise in my previous papers. The purpose of this journal is relates to re-organizing, comparing and analyzing my researches that I did. Also, it will shows the effectiveness of the deep learning algorithms as a kind of a radar detection method.
    [4] Jihoon Kwon and Nojun Kwak, "Application of Deep Neural Networks for UWB Radar ", - , 2018
    This paper will propose the deep learning classification methods for UWB radar. This research is going to be connected with the results of the paper by J. Bryan, J. Kwon, N. Lee and Y. Kim "Application of UWB radar for classification of human activities."
    [5] Jihoon Kwon and Nojun Kwak, "Recurrent Neural Networks for Recognizing the Micro-Doppler Signals by Human Walking and Background Noise", International Journal of Antennas and Propagation, 2018
    Doppler signals are time series signals. So, We expect to be able to apply Recurrent Neural Networks (RNN) as a methodology to recognize the micro-Doppler signals. We expect RNN to provide interesting results.

Domestic Journals
    [1] 권지훈, 곽노준 "복잡한 클러터 환경에서의 RANSAC 알고리즘을 이용한 레이더 초기화 기법에 대한 연구", 한국전자파학회논문지, 2018.