Improved Clutter Suppression Method for Bridge Vibration Measurement Based on GB-MIMO Radar

Ground based multi-input multi-output (GB-MIMO) radar has been gradually applied in the field of bridge vibration monitoring. During measurement, the interference of static clutter on the vibrating target signal in the same azimuth and range resolution severely affects the vibration measurement. The traditional method of removing static clutter is to estimate static clutter by fitting a circular curve to a complex signal point set, that is, fitting the vector of the circular curve center deviating from the origin of the polar coordinate system. Due to the clustered distribution of vibration points during non vibration time, the trajectory of complex sequence point sets will be distorted, resulting in poor clutter suppression effect. This article proposes an improved method for measuring the dynamic deflection of bridges by suppressing clutter. Firstly, obtain the vibration time period of the vibration candidate points through time-frequency analysis. Secondly, extract the complex sequence of each pixel during the vibration time period, and use the circular trace fitting method to estimate the stationary clutter. Then, based on outlier degree and outlier rate evaluation indicators, select vibration measurement pixels that can achieve effective clutter suppression. The effectiveness of this method was preliminarily verified by comparing the dynamic deflection curves of millimeter wave radar and laser vibrometer through model bridge experiments. In addition, the method was also validated through a high-speed railway bridge experiment.