Improved Radar Localization on Lidar Maps Using Shared Embedding
This work addresses localization challenges for autonomous vehicles by enabling radar-based navigation on existing lidar maps, though it appears incremental as it builds on prior methods like RaLL.
The paper tackles the problem of radar global localization and pose tracking on pre-built lidar maps by using deep neural networks to create a shared embedding space, improving similarity measurement for map retrieval and data matching. It demonstrates effectiveness on RobotCar and MulRan datasets with comparisons to Scan Context and RaLL, and reduces neural networks in the pose tracking pipeline compared to RaLL.
We present a heterogeneous localization framework for solving radar global localization and pose tracking on pre-built lidar maps. To bridge the gap of sensing modalities, deep neural networks are constructed to create shared embedding space for radar scans and lidar maps. Herein learned feature embeddings are supportive for similarity measurement, thus improving map retrieval and data matching respectively. In RobotCar and MulRan datasets, we demonstrate the effectiveness of the proposed framework with the comparison to Scan Context and RaLL. In addition, the proposed pose tracking pipeline is with less neural networks compared to the original RaLL.