Shaojun Xu

Zhejiang University, Tsinghua University

Papers

Total Citations

H-Index

About

Shaojun Xu is pioneering the intersection of natural language processing and formal task specification for multi-robot systems. Their core research addresses a critical bottleneck in robotics: enabling non-experts to intuitively command complex, long-horizon collaborative missions. Xu’s foundational work, "Decomposition-Based Hierarchical Task Allocation and Planning for Multi-Robots Under Hierarchical Temporal Logic Specifications" (2024, 13 citations), directly tackles the scalability issues of traditional Linear Temporal Logic (LTL) methods. By introducing a hierarchical framework, Xu’s approach simplifies the specification of intricate tasks that would otherwise result in unwieldy, single-formula LTL expressions, thereby making multi-robot planning more interpretable and computationally tractable. Building on this, their latest contribution, "Nl2Hltl2Plan" (2025), leverages large language models to translate natural language commands into these hierarchical specifications, addressing the critical challenge of translation ambiguity. This work is a significant step toward democratizing multi-robot control, allowing users without formal logic training to deploy and manage complex robotic teams. Xu’s research is not only technically rigorous but also practically impactful, directly bridging the gap between human intent and autonomous robotic execution.