Four-Channel Phase and Quadrature Self-Injection-Locked (PQSIL) Radar for Displacement Monitoring Using a Modified Principal Component Analysis (MPCA) Method
Ji-Xun Zhong, Ju-Yin Shih, Fu-Kang Wang
Abstract
First, this study proposed a four-channel phase and quadrature self-injection-locked (PQSIL) radar by continuously switching four quadrature injection phases. Thereafter, a modified principal component analysis algorithm was proposed and applied on the four channels, and the measured result was confirmed to overcome the nonlinear distortion induced by the self-injection-locked (SIL) mechanism and accurately reflect the information of a moving target. In addition, using a series of the proposed elliptical trajectory calibration, the clutter effect can be canceled and the target’s displacement can be accurately recovered. A closed-loop experiment was conducted to verify that when the radar was used for displacement monitoring, the clutter signal can change and affect the Doppler signal, which causes distortion in the recovered phase and a dynamic range limitation for the traditional SIL radar. In this study, the four-channel PQSIL radar was operated at 2.4 GHz to monitor a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$10\times10$ </tex-math></inline-formula> cm2 metal plate moved by an actuator and with a <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$30\times30$ </tex-math></inline-formula> cm2 static metal plate near the actuator. The sensing displacement could be measured from <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$0.125\lambda $ </tex-math></inline-formula> to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1\lambda $ </tex-math></inline-formula> at a distance of 3 m, with root-mean-square errors (RMSEs) of less than 1%.