Assessing Important Uncertainty Influences of Ground-Based Radar for Bridge Monitoring

Remote sensors such as ground-based interferometric radar (GBR) are increasingly applied in bridge monitoring to determine damage-sensitive features. A crucial part of successfully determining these features is the removal of multiple influences, which can cause inaccurate measurements. This study highlights two aspects, which predominantly determine the measurement accuracy: clutter estimation and the projection from measurements in line of sight (LOS) to a Cartesian coordinate system. Furthermore, the removal of these influences can also introduce additional uncertainty, which is rarely considered. The uncertainty is estimated exemplarily for measurements at a railway bridge in Karlsruhe (Germany) and compared to the results of measurements from a terrestrial laser scanner (TLS). Without consideration of this uncertainty, the GBR would compare favorably to the TLS even though both sensors achieve a similar measurement uncertainty in the order of 0.1-0.3 mm. Our approach for uncertainty estimation could be further developed with a simulation of the signal footprint to approximate its size better.