Caizhen He

Papers

Total Citations

H-Index

About

Caizhen He is a researcher advancing the field of trajectory prediction, with a focus on enabling safer autonomous systems and socially aware robotics. His work addresses the fundamental challenge of forecasting pedestrian motion in dynamic, interactive environments, where human-human and human-object interactions introduce significant uncertainty. He introduced TPPO (Trajectory Predictor with Pseudo Oracle) in 2024, a novel approach that has already garnered 31 citations for its innovative handling of future uncertainties through pseudo-oracle guidance. Earlier, he developed an adversarial learned trajectory predictor incorporating knowledge-rich latent variables (2020), demonstrating his sustained commitment to improving prediction robustness. He’s contributions are critical for applications ranging from autonomous driving to crowd-aware navigation, where accurate forecasting directly impacts safety and efficiency. By tackling the complexities of human randomness and social interactions, He is helping to bridge the gap between current prediction models and the reliability required for real-world deployment.