Papers
13
Total Citations
366
H-Index
9
About
Mingyu Fan is a prolific robotics and artificial intelligence researcher whose work spans trajectory prediction, autonomous navigation, human-robot interaction, and bio-inspired robotics. He is perhaps best known for his contributions to pedestrian trajectory forecasting, where his attention-based spatio-temporal graph neural network, AST-GNN, has garnered nearly 200 citations since its 2021 publication — establishing him as a leading voice in modeling complex social interactions within crowded environments. His subsequent works, including Tra2Tra and CSR, further refined trajectory prediction through global spatial-temporal attention mechanisms and variational autoencoding, collectively advancing safe autonomous robot navigation. Beyond trajectory forecasting, Fan has demonstrated impressive breadth: his bio-inspired soft robotic system for deep-sea exploration, drawing from deep-sea snail locomotion using shape memory alloys, has attracted 27 citations since 2024 and exemplifies his creativity at the intersection of biology and engineering. His more recent research addresses intelligent human-robot interaction through large language models, enabling zero-shot voice and posture-based communication — work especially relevant for aging societies. Fan also contributes to UAV path planning and lifelong robotic learning, reflecting a research philosophy that bridges theoretical machine learning with real-world autonomous systems applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5UAV Path Planning with an Adaptive Hybrid PSO15 citations · 2023
- 6
- 7
- 8
- 9
- 10Online Active Continual Learning for Robotic Lifelong Object Recognition7 citations · 2023