Yusheng Peng
Papers
2
Total Citations
27
H-Index
2
About
Yusheng Peng is a researcher at the forefront of computer vision and autonomous systems, specializing in human trajectory prediction—a critical technology for intelligent surveillance, robot navigation, and autonomous driving. Peng’s major contributions lie in advancing the accuracy and efficiency of forecasting pedestrian movements. In their highly cited 2023 work, "MRGTraj," Peng introduced a novel non-autoregressive approach that overcomes the cumulative error problem inherent in traditional RNN- and Transformer-based autoregressive models, achieving 23 citations and setting a new standard for trajectory prediction. Earlier, in "PECGAN" (2021), Peng pioneered an endpoint-conditioned generative adversarial network that decodes motion features without self-recurrent architecture, demonstrating innovative thinking in leveraging GANs for spatial-temporal forecasting. With a combined citation impact of 27 citations, Peng’s work is recognized for addressing real-world challenges in dynamic environments. Their research not only pushes the boundaries of predictive modeling but also directly enhances the safety and reliability of autonomous systems, making Yusheng Peng a notable emerging voice in the intersection of deep learning and intelligent transportation.
Research Focus
Key Achievements
Top Papers
- 1MRGTraj: A Novel Non-Autoregressive Approach for Human Trajectory Prediction23 citations · 2023
- 2