This paper proposes a YOLO(You Only Look Once)-based method for simultaneously estimating the number, distances, and velocities of multiple targets in an OFDM(orthogonal frequency division multiplexing) radar environment. The proposed method detects the peak region corresponding to each target in a two-dimensional periodogram and estimates the number of targets and their distance–velocity values using the number and center coordinates of the detected regions. By employing a single-stage object detection architecture, the proposed method provides a simpler estimation procedure than conventional CNN(convolutional neural network)-based methods that require separate regression models according to the number of targets. Averaged over all test conditions with one to five targets and SNRs ranging from -10 to 20 dB, the proposed method achieves a velocity mean absolute error of 1.34 km/h and a distance mean absolute error of 0.71 m with 32 OFDM symbols, outperforming the CA-CFAR(cell-averaging constant false alarm rate) and CNN-based methods. The proposed method also maintains relatively stable distance and velocity estimation performance as the number of targets increases.
한국어
본 연구에서는 OFDM(orthogonal frequency division multiplexing) 레이다 환경에서 YOLO(You Only Look Once)를 활용하여 다중 표적의 개수, 속도 및 거리를 동시에 추정하는 기법을 제안한다. 제안 기법은 2차원 주기도에서 각 표적의 피크 영역을 검출하고, 검출된 영역의 개수와 중심좌표를 이용하여 표적의 개수와 거리·속도를 추정한다. 단일 스테이지 객체 검출 구조를 사용하므로 표적 수에 따라 별도의 회귀 모델을 적용하는 기존 CNN(convolutional neural network) 기반 기법보다 추정 절차가 간결하다. 모의실험 결과, 표적 수 1~5 개와 SNR -10~20 dB의 전체 시험 조건에서 조건별 MAE를 평균하였을 때, OFDM 심볼 개수가 32개인 경우 속도 평균절대오차 1.34 km/h와 거리 평균절대오차 0.71 m를 기록하였으며, CA-CFAR(cell-averaging constant false alarm rate) 및 CNN 기반 기법보다 낮은 추정 오차를 보였다. 또한 표적 수가 증가하는 환경에서도 상대적으로 안정적인 거리 및 속도 추정 성능을 유지하였다.
목차
요약 Abstract Ⅰ. 서론 Ⅱ. OFDM 레이다 시스템 Ⅲ. 2차원 주기도 Ⅳ. 표적 추정 기법 4.1 기존 CA-CFAR 4.2 기존 CNN 4.3 제안하는 YOLO Ⅴ. 모의실험 5.1 모의실험 환경 5.2 모의실험 결과 Ⅵ. 결론 REFERENCES
Ever since next generation convergence technology became one of the most important industries in the nation, computing professionals have encountered a growing number of challenges. Along with scholars and colleagues in related fields, they have gathered in avariety of forums and meetings over the last few decades to share their knowledge, experiences and the outcome of their research. These exchanges have led to the founding of the International Next-generation Convergence technology (INCA) on December 1, 2015. INCA was registered as an incorporated association under the Ministry of Information and Communications. The main purpose of the organization is to improve our society by achieving the highest capability possible in next generation convergence technology.
간행물
간행물명
차세대융합기술학회논문지 [The Journal of Next-generation Convergence Technology Association]