Fei Xiaoxiao

Papers

Total Citations

H-Index

About

Fei Xiaoxiao is a roboticist whose research lies at the intersection of autonomous navigation, visual place recognition, and map-based localization. Her work addresses a fundamental challenge in robotics: how can a mobile robot efficiently and accurately recognize where it is, even in large, unstructured environments? Xiaoxiao’s major contributions include pioneering grammar-based map parsing for view-invariant map retrieval, a technique that enables robust similarity search over large collections of 2D pointset maps—a critical capability for long-term autonomous operation. Her most cited paper (2017, 4 citations) introduces this approach as a powerful alternative to traditional bag-of-words methods. Additionally, her work on unsupervised place discovery (2016, 3 citations; 2017, 2 citations) explores how deep convolutional neural networks (DCNNs) can be leveraged to automatically partition a robot’s workspace into meaningful places, optimizing classification accuracy, precision, and recall without requiring manual labeling. This unsupervised paradigm is particularly impactful for robots operating in unknown or dynamic environments. Though early in her career, Xiaoxiao’s contributions are foundational for scalable, intelligent robotic mapping and localization systems.