Document Type : Original Article
Lecturer, Department of Electrical and Computer Engineering Military Technical College, Kobry El-Koba, Cairo, Egypt.
The purpose of moving target indicator (MTI) radars is to reject signals from fixed unwanted targets, such as buildings, hills, and trees, and retain signals for detection of moving targets such as aircraft, ships, missiles,..etc. There are many conventional MTI techniques such as, the adaptive pulse canceller, doppler filter banks, and fast Fourier transform (FFT) filter banks. Each technique suffers certain considerable limitation. This paper utilizes a neural network approach to implement a system that can be used to detect moving targets in severely cluttered environment. A three-layer neural network has been trained to perform as moving target indicator. The back-propagation training algorithm has been used to train the proposed network. Significant performance improvements have been achieved as compared with the conventional fast Fourier transform method.