Publications

2026

1. 

   AURA: Asymptotically Optimal Uncertainty-Robust Replanning Algorithm for Kinodynamic Systems

   Seyedali Golestaneh, Zhuoyun Zhong, Donghyung Lee, and Constantinos Chamzas

   IEEE Robotics and Automation Letters, 2026

   Accepted

   Abs arXiv Bib PDF Video Code

   AURA combines asymptotically optimal kinodynamic planning with online replanning and control refinement. Simulation and real-world evaluations show improved trajectory quality and tracking under motion uncertainty.

   @article{golestaneh2026aura,
     title = {AURA: Asymptotically Optimal Uncertainty-Robust Replanning Algorithm for Kinodynamic Systems},
     author = {Golestaneh, Seyedali and Zhong, Zhuoyun and Lee, Donghyung and Chamzas, Constantinos},
     journal = {IEEE Robotics and Automation Letters},
     year = {2026},
     archiveprefix = {arXiv},
     url = {https://arxiv.org/abs/2605.27699},
     note = {Accepted},
   }

2. 

   MetaPusher: Meta Learning and Planning for Nonprehensile Manipulation of Unseen Objects with Rapid Online Adaptation

   Donghyung Lee, Seyedali Golestaneh, Jaskrit Singh, Zhuoyun Zhong, Athanasios Kapoutsis, and Constantinos Chamzas

   2026

   Abs arXiv Bib PDF

   MetaPusher meta-learns manipulation dynamics for unseen objects and adapts its model online while reusing and refining a kinodynamic planner’s search tree. The framework is evaluated in simulation and sim-to-real tasks.

   @misc{golestaneh2026metapusher,
     title = {MetaPusher: Meta Learning and Planning for Nonprehensile Manipulation of Unseen Objects with Rapid Online Adaptation},
     author = {Lee, Donghyung and Golestaneh, Seyedali and Singh, Jaskrit and Zhong, Zhuoyun and Kapoutsis, Athanasios and Chamzas, Constantinos},
     year = {2026},
     archiveprefix = {arXiv},
     url = {https://arxiv.org/abs/2609.21122},
   }

3. 

   Terminal Matters: Kinodynamic Planning with a Terminal Cost and Learned Uncertainty in Belief State-Cost Space

   Zhuoyun Zhong, Seyedali Golestaneh, and Constantinos Chamzas

   2026

   Submitted to IEEE Transactions on Robotics

   Abs arXiv Bib PDF Video Code Website

   KiTe adds a terminal-cost objective to asymptotically optimal kinodynamic planning and extends it to belief space. It learns dynamics and uncertainty from data and evaluates goal-reaching performance in simulation and on physical systems.

   @misc{zhong2026kite,
     title = {Terminal Matters: Kinodynamic Planning with a Terminal Cost and Learned Uncertainty in Belief State-Cost Space},
     author = {Zhong, Zhuoyun and Golestaneh, Seyedali and Chamzas, Constantinos},
     year = {2026},
     archiveprefix = {arXiv},
     url = {https://arxiv.org/abs/2605.09046},
     note = {Submitted to IEEE Transactions on Robotics},
   }

4. 

   CoAd: Constant-Time Planning for Continuous Goal Manipulation with Compressed Library and Online Adaptation

   Adil Shiyas^*, Zhuoyun Zhong^*, and Constantinos Chamzas

   2026

   Submitted to ICRA 2027

   Abs arXiv Bib PDF Video Code Website

   CoAd provides constant-time planning over continuous goal-parameterized manipulation tasks using a compressed motion library and lightweight online adaptation. The approach is evaluated across simulated and real-world manipulation tasks.

   @misc{shiyas2026coad,
     title = {CoAd: Constant-Time Planning for Continuous Goal Manipulation with Compressed Library and Online Adaptation},
     author = {Shiyas, Adil and Zhong, Zhuoyun and Chamzas, Constantinos},
     year = {2026},
     archiveprefix = {arXiv},
     url = {https://arxiv.org/abs/2603.12488},
     note = {Submitted to ICRA 2027},
   }

5. 

   ActivePusher: Active Learning and Planning with Residual Physics for Nonprehensile Manipulation

   Zhuoyun Zhong, Seyedali Golestaneh, and Constantinos Chamzas

   In IEEE International Conference on Robotics and Automation, 2026

   Best Student Paper at ICRA 2026 and HRF 2026; ICRA Planning and Control Best Paper finalist

   Abs arXiv Bib PDF Video Code Website

   ActivePusher combines residual-physics dynamics learning with uncertainty-guided active data collection and planning for nonprehensile manipulation. Experiments in simulation and on a physical robot show improved data efficiency and planning success.

   @inproceedings{zhong2026activepusher,
     title = {ActivePusher: Active Learning and Planning with Residual Physics for Nonprehensile Manipulation},
     author = {Zhong, Zhuoyun and Golestaneh, Seyedali and Chamzas, Constantinos},
     booktitle = {IEEE International Conference on Robotics and Automation},
     year = {2026},
     archiveprefix = {arXiv},
     url = {https://arxiv.org/abs/2506.04646},
     note = {Best Student Paper at ICRA 2026 and HRF 2026; ICRA Planning and Control Best Paper finalist},
   }

2024

1. 

   Expansion-GRR: Efficient Generation of Smooth Global Redundancy Resolution Roadmaps

   Zhuoyun Zhong, Zhi Li, and Constantinos Chamzas

   In IEEE/RSJ International Conference on Intelligent Robots and Systems, 2024

   Abs arXiv Bib PDF Video Code Website

   Expansion-GRR uses configuration-space projections and continuity-aware expansion to generate smooth global redundancy-resolution roadmaps efficiently. Simulated and real-arm teleoperation experiments show faster roadmap generation and improved solution quality.

   @inproceedings{zhong2024expansiongrr,
     title = {Expansion-GRR: Efficient Generation of Smooth Global Redundancy Resolution Roadmaps},
     author = {Zhong, Zhuoyun and Li, Zhi and Chamzas, Constantinos},
     booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
     year = {2024},
     archiveprefix = {arXiv},
     url = {https://ieeexplore.ieee.org/document/10801917},
   }

2023

1. 

   A Shared Autonomous Nursing Robot Assistant with Dynamic Workspace for Versatile Mobile Manipulation

   Nikita Boguslavskii^*, Zhuoyun Zhong^*, Lorena Maria Genua^*, and Zhi Li

   In IEEE/RSJ International Conference on Intelligent Robots and Systems, 2023

   * Equal contribution

   Abs DOI Bib Code

   This work presents a shared-autonomy nursing robot assistant with a dynamically managed workspace to support versatile mobile manipulation in care environments.

   @inproceedings{boguslavskii2023nursingrobot,
     title = {A Shared Autonomous Nursing Robot Assistant with Dynamic Workspace for Versatile Mobile Manipulation},
     author = {Boguslavskii, Nikita and Zhong, Zhuoyun and Genua, Lorena Maria and Li, Zhi},
     booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems},
     year = {2023},
     url = {https://ieeexplore.ieee.org/document/10342401},
     doi = {10.1109/IROS55552.2023.10342401},
     note = {* Equal contribution}
   }

2. 

   Self-Supervised Pre-Training for Robust and Generic Spatial-Temporal Representations

   Mingzhi Hu, Zhuoyun Zhong, Xin Zhang, Yanhua Li, Yiqun Xie, Xiaowei Jia, Xun Zhou, and Jun Luo

   In IEEE International Conference on Data Mining, 2023

   Abs DOI Bib

   This paper introduces a self-supervised pre-training method for learning robust, general-purpose spatial-temporal representations and evaluates their transfer to downstream prediction tasks.

   @inproceedings{hu2023selfsupervised,
     title = {Self-Supervised Pre-Training for Robust and Generic Spatial-Temporal Representations},
     author = {Hu, Mingzhi and Zhong, Zhuoyun and Zhang, Xin and Li, Yanhua and Xie, Yiqun and Jia, Xiaowei and Zhou, Xun and Luo, Jun},
     booktitle = {IEEE International Conference on Data Mining},
     year = {2023},
     url = {https://ieeexplore.ieee.org/document/10415669},
     doi = {10.1109/ICDM58522.2023.00024},
   }