Edge‐Enabled DSRC Deployment in Highway Scenarios for Vehicle Lane Positioning and Driving Patterns Tracking

Accurate lane‐level positioning and vehicle manoeuvre tracking are essential for modern highway applications, such as toll auditing and trajectory verification, and higher level connected and automated vehicle applications. However, on‐board GNSS/INS sensing degrades during manoeuvres or occlusions and reduces the positioning reliability. To address these challenges, this study proposes an edge‐enabled positioning framework that fuses on‐board GNSS/INS measurements with roadside sensing information to achieve high‐precision, real‐time estimation of vehicle states. The proposed framework combines vehicle‐based observations in the longitudinal‐lateral coordinate and roadside measurements in polar form. The typical vehicle motions, such as acceleration, deceleration, and lane‐changing, are represented using multiple state‐space models (CA, CV, and CT), and tracking is performed by parallel linear (KF) and nonlinear filters integrated via an interacting multiple model algorithm to enhance robustness under manoeuvring conditions. Field experiment results demonstrate that the proposed edge‐assisted fusion strategy substantially improves lane‐level trajectory accuracy and reliability, particularly when vehicles exhibit rapid or unstable manoeuvres, thereby supporting practical vehicle‐road cooperative applications.