Open AccessCNN based Distance and Velocity Estimation for OFDM Radar Systems
Author(s) -
Jinseok Choi,
Eui-Rim Jeong
Publication year - 2022
Publication title -
webology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.259
H-Index - 18
ISSN - 1735-188X
DOI - 10.14704/web/v19i1/web19313
Subject(s) - estimator , radar , orthogonal frequency division multiplexing , computer science , algorithm , fast fourier transform , periodogram , noise (video) , mathematics , telecommunications , artificial intelligence , statistics , image (mathematics)
The objective of this paper is to propose a new target distance and velocity estimation technique for OFDM radar systems. First, the 2D periodogram is collected from the reflected signal via FFT of received OFDM symbols. The largest value of a 2D periodogram often represents the target so that its position indicates the distance and velocity. The CFAR is one of the famous conventional techniques to find the peak in the 2D periodogram. In this paper, a CNN based estimator is proposed. The proposed CNN directly finds the distance and velocity from the 2D periodogram. The proposed method requires only 2D periodogram to estimate the target’s distance and velocity. On the other hand, the conventional methods need noise variance as well as the periodogram. The performance is examined through computer simulation. In the simulation, the MAEs are compared between the conventional and proposed methods. According to the results, the MAEs of the proposed method are lower approximately 8 m in distance and 7 km/h in speed to the conventional method. The proposed OFDM radar technique can be applied to 6G mobile communications to identify the moving targets without additional frequency resource allocation for the radar system. In other words, by using the proposed technique, the convergence of the communication and radar can be possible.