Author(s)

Zhiyuan Wen

Corresponding Author

Zhiyuan Wen

Abstract

This paper systematically compares the technical routes of Waymo, Tesla, and academic research in autonomous driving environmental perception and decision-making. Waymo adopts a LiDAR-centric multi-sensor fusion approach targeting L4 Robotaxi markets, while Tesla employs an end-to-end pure vision architecture for consumer-grade FSD systems. Academic research focuses on lightweight models and trustworthy decision frameworks. Analysis reveals that technical divergences stem from sensor configurations, data capabilities, and commercialization strategies, with future trends leaning toward cost-effective multimodal fusion and edge-cloud collaboration.

Keywords

Autonomous Driving; Environmental Perception; Multimodal Fusion; End-to-End Learning; Technical Roadmap