Robust Modified MVDR Scheme Using Chirp Signal for Direction of Arrival Estimation

Abstract

This research is about an effort to increase the robustness of the Minimum Variance Distortionless Response (MVDR) algorithm to noise by using a chirp signal for direction of arrival estimation (DoA). DoA is a part of radar capability to estimate the angle of arrival on the object under observation. The conventional MVDR as proposed by J. Capon, was designed to work with the monochromatic sinusoidal signal. Even though the conventional MVDR work on low SNR up to 0 dB, however, the conventional method does not work well if chirp signal is used instead of monochromatic sinusoidal signal. The usage of MVDR chirp signal is essential in the case of a very low SNR environment such as in long distance object detection, which is typically more than 10 km. The problem to be solved in this research is how to modify the MVDR algorithm so that it can work well on chirp signal. In this research we offer a modified MVDR algorithm by adding the matched filter and the phase detector components before the MVDR algorithm is applied. Matched filter is responsible for the timing of the chirp signal detection, and the phase detector is to estimate the time delay estimation of each chirp signal from each antenna with a reference signal, which correspond to the phases. Based on the phase estimation, sinusoidal signal is generated and fed to the MVDR algorithm. On the technical aspect, the chirp signal is sent intermittently with a duration of 100 ?s and repeated in time interval of 1 ms. The antenna sensor using an array of Uniform Linear Array (ULA) which consist of N-elements. Computer simulation shows that the modified MVDR using the chirp signal improve the robustness of the algorithm up to -30 dB, while on the other hand the classical MVDR works only up to 0 dB SNR. -30 dB of SNR is the minimum requirement of 3D Radar existing.