Robust Localization for Autonomous Vehicles in Highway Scenes

A new research paper from Daqian Cheng and colleagues presents a robust localization system specifically designed for autonomous vehicles on highways, an area that has been under-explored compared to urban environments. The system tackles key highway challenges like environment changes under information homogeneity, heavy occlusion, and degraded GNSS signals. It features a dual-likelihood LiDAR front end that decouples 3D geometric structures and 2D road-texture cues to handle environmental shifts, and a Control-EKF that uses steering and acceleration commands to reduce lag and improve closed-loop behavior. The system also includes an automated offline mapping and ground-truth pipeline to keep maps fresh at high cadence.

Benchmarked against industry standards Apollo and Autoware, the proposed system performs similarly on urban roads but shows superior robustness on challenging highway scenarios. It has been validated by over one million kilometers of road testing. To catalyze further progress, the authors release a public dataset covering both urban roads and highways, totaling 163 km, with a focus on representative challenging highway clips. The paper has been accepted to the 2026 IEEE International Conference on Robotics and Automation (ICRA), signaling its significance in the autonomous driving community.