Cheng Chi

Shanghai Jiao Tong University, Beijing Academy of Artificial Intelligence

Papers

Total Citations

H-Index

About

Cheng Chi is a researcher working at the intersection of autonomous navigation and robotic intelligence, with particular expertise in sensor fusion, state estimation, and failure-aware robotic systems. His most prominent contribution is GICI-LIB, an open-source GNSS/INS/Camera integrated navigation library that has rapidly gained traction in the robotics and autonomous driving communities, accumulating over 40 citations since its 2023 release. This work addresses a critical challenge in real-world autonomy: achieving robust, high-accuracy localization across diverse and challenging environments by tightly coupling complementary sensing modalities. The library has become a valuable resource for researchers and engineers developing navigation solutions for autonomous vehicles and mobile robots alike. More recently, Chi has expanded his research into the domain of robotic failure detection with Code-as-Monitor (CaM), a novel framework that leverages visual programming to enable both reactive and proactive identification of open-set failures in closed-loop robotic systems. This work reflects a growing sophistication in his research agenda, moving from low-level state estimation toward higher-level robot reliability and autonomy. Together, his contributions demonstrate a coherent vision: building autonomous systems that are not only accurate in perception and navigation, but also resilient and self-aware in the face of real-world uncertainty.