| Title |
Convolutional Neural Network-based Target Detection Method for Passive Bistatic Radar using FM Broadcasting Signals |
| Authors |
박근호(Geun-Ho Park) ; 김형남(Hyoung-Nam Kim) |
| DOI |
https://doi.org/10.5573/ieie.2020.57.12.70 |
| Keywords |
Passive bistatic radar; Constant false alarm rate detector; Convolutional neural network; Target detection |
| Abstract |
Frequency-modulation (FM) band passive bistatic radar is a system for estimating the position and speed of a fast-moving target. Such a system can exploit commercial FM-radio broadcasting signals in the frequency band of 88-108 MHz. Most radar systems, including passive bistatic radars, perform target detection using a constant false alarm rate (CFAR) algorithm. However, because of the characteristics of FM broadcasting signals, many sidelobes appear in the range-Doppler map, and the CFAR detection algorithm produces unwanted detection results. Besides, as the shape of the ambiguity function changes over time, this property makes it difficult to detect the target. This paper proposes a target detection method based on a convolutional neural network that can solve this problem. This method can achieve almost the identical detection probability as a CFAR detector but has a small number of false detections generated by sidelobes. We propose a method of training a convolutional neural network for target detection and compare the performances of the proposed algorithm and a CFAR detector through computer simulations. |