Shaoheng Fang

Papers

Total Citations

H-Index

About

Shaoheng Fang is a rising researcher in autonomous driving and robotics, whose work centers on self-supervised learning for bird’s eye view (BEV) motion prediction and cross-modality signal integration. His most notable contribution, the 2024 paper “Self-Supervised Bird’s Eye View Motion Prediction with Cross-Modality Signals,” tackles a critical challenge in the field: learning dense BEV motion flow without costly manual annotations. Fang identified that existing self-supervised methods, which depend on point cloud correspondences, often suffer from “fake flow” and temporal inconsistency. To address this, he proposed a novel framework that leverages cross-modality signals—combining visual and LiDAR data—to generate more accurate and robust motion predictions. This work has already garnered 5 citations within its first year, signaling strong early impact and relevance. Fang’s research is particularly significant for advancing perception systems in autonomous vehicles, where reliable motion understanding is essential for safe navigation. By pioneering self-supervised techniques that mitigate data inefficiencies, he is helping to reduce reliance on expensive labeled datasets, making autonomous systems more scalable. For students and researchers, Fang’s work exemplifies how innovative cross-modal learning can solve persistent problems in real-world robotics, offering a promising direction for future exploration in BEV perception and motion forecasting.