Moji Shi

Shanghai Artificial Intelligence Laboratory, Delft University of Technology

Papers

Total Citations

H-Index

About

Moji Shi is a robotics researcher whose work lies at the intersection of humanoid locomotion and multi-agent autonomy. His primary contributions span two challenging domains: enabling stable, perception-driven walking for humanoid robots, and developing scalable collision avoidance for swarms of micro aerial vehicles (MAVs). In his highly cited 2025 paper on humanoid locomotion, Shi introduced a perceptive internal model that addresses the critical instability of bipedal robots—a problem where even minor perceptual delays can cause falls. This work, already garnering 7 citations, contrasts with simpler “blind” quadruped policies by integrating real-time visual feedback. For multi-agent systems, his 2024 framework leverages spatiotemporal occupancy grid maps (SOGM) to predict future obstacle positions, enabling decentralized trajectory planning that handles both static and dynamic threats. This paper has earned 5 citations for its practical impact on drone swarm coordination. Shi also contributed to benchmarking standards by proposing metrics to quantify environmental difficulty for dynamic obstacle avoidance, a foundational tool for fair algorithm comparison. His research is notable for bridging theoretical control challenges with real-world deployment constraints, making him a rising figure in autonomous robotics.