Predictive modeling and evaluation of LiDAR sensors through automotive windshields with anti-reflective coatings

Abstract

The autonomous vehicle (AV) necessitates accurate environmental perception and robust localization. Among the various sensing technologies, Light Detection and Ranging (LiDAR) demonstrates exceptional spatial resolution and distance measurement capabilities. Despite these advantages, external LiDAR placement often leads to aerodynamic drag, aesthetic concerns, and exposure to harsh conditions.S Mounting LiDAR behind the windshield enhances vehicle design and operational efficiency, but introduces new challenges related to the windshield's optical properties, including signal attenuation and distortion, amplified by theS double transmission of the laser beam through the glass. Anti-reflective (AR) coatings minimize reflections at critical interfaces, thus preserving signal strength to offer a promising solution. However, the lack of a standardized evaluation methodology complicates the determination of an optimal AR coating recipe. To address this lack, this study introduces a LiDAR signal strength prediction model that employs Fresnel equations to account for thickness, refractive index, curvature, and round-trip path effects. A comprehensive validation protocol compares empirical measurements under different windshield conditions, and includes a thermal aging test. This work provides a roadmap for systematic sensor calibration, streamlining LiDAR integration in next-generation AVs and improving measurement reliability. © 2025 Elsevier B.V., All rights reserved.