LiDAR localization at 100 FPS: A map-aided and template descriptor-based global method

Pengcheng Shi, Jiayuan Li, Yongjun Zhang

Abstract: Highlights•We propose a template descriptor based global localization method, which can run at 100 FPS on a single-thread CPU.•We propose a template descriptor with specific loss function to guarantee the localization efficiency and accuracy.•We design a candidate points generation method and use database management to reduce the pressures of map maintenance.•We design the experiments with both public and self-collected datasets to evaluate our localization performance.