We present a robust and precise localization system that achievescentimeter-level localization accuracy in disparate city scenes. Our systemadaptively uses information from complementary sensors such as GNSS, LiDAR, andIMU to achieve high localization accuracy and resilience in challenging scenes,such as urban downtown, highways, and tunnels. Rather than relying only onLiDAR intensity or 3D geometry, we make innovative use of LiDAR intensity andaltitude cues to significantly improve localization system accuracy androbustness. Our GNSS RTK module utilizes the help of the multi-sensor fusionframework and achieves a better ambiguity resolution success rate. Anerror-state Kalman filter is applied to fuse the localization measurements fromdifferent sources with novel uncertainty estimation. We validate, in detail,the effectiveness of our approaches, achieving 5-10cm RMS accuracy andoutperforming previous state-of-the-art systems. Importantly, our system, whiledeployed in a large autonomous driving fleet, made our vehicles fullyautonomous in crowded city streets despite road construction that occurred fromtime to time. A dataset including more than 60 km real traffic driving invarious urban roads is used to comprehensively test our system.
展开▼
