Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces OLRA, a low-cost framework for lane-based map localization that aligns driver-view routes with camera-detected lane markings to enhance both vehicle localization accuracy and visual route consistency. By establishing practical evaluation metrics and benchmarking against OpenPilot, OLRA demonstrates superior performance in complex road segments and at distances over 20 meters, achieving a lower overall Euclidean error. The findings suggest a promising direction for developing intuitive driving guidance systems that are both effective and economical.
OLRA outperforms existing methods by achieving lower localization errors in complex driving scenarios, paving the way for more intuitive navigation systems.
Driver-centric route representation plays a vital role in intuitive driving guidance systems. This paper presents OLRA, a low-cost, map-localization-based framework that derives driver-view-aligned routes by matching map-based navigation routes with camera-detected lane markings. This alignment process mutually enhances vehicle localization accuracy and visual route consistency. To bridge the evaluation gap across different paradigms, we introduce practical route evaluation metrics and benchmark OLRA against OpenPilot, a representative direct-generation approach. Experimental results on the nuScenes dataset demonstrate that OLRA outperforms OpenPilot in complex road segments and in route estimation at distance beyond 20 meters, achieving lower overall Euclidean error. This study is expected to promote future research in low-cost, maplocalization-based route generation methods.