Localization of Embedded Sensors in Reinforced Concrete via Time-Series Magnetic Field Sensing and Maximum Likelihood Estimation

To facilitate nonintrusive inspection of aging road infrastructure, we propose a framework in which sensors are embedded in reinforced concrete during construction and later interrogated by a passing inspection vehicle. Precise localization of these sensors, despite limited power and time windows, is crucial to realize this concept. This letter introduces a localization method that exploits time-series magnetic-field signatures captured as a moving anchor passes overhead. Coarse positions are first derived from sequential sensor data, then iteratively refined through maximum-likelihood estimation with distance-weighted confidence. Experiments indicate that the proposed method reduces localization error by 23%–43% relative to conventional techniques, such as simple averaging or nearest-anchor-based estimation, even when the minimum anchor-sensor gap reaches 1 m, underscoring its promise for drive-by infrastructure monitoring without fixed anchors or lane closures.