Zirui (Ray) Xu

I'm a Ph.D. candidate in Aerospace Engineering at University of Michigan, advised by Vasileios Tzoumas. My goal is to realize scalable and reliable coordination in resource-constrained, unstructured, and untrustworthy environments, such as smart cities, environmental monitoring, and disaster response. I develop and leverage tools of optimization, control, and robotics to provably enable self-(re)configurable coordination for fully distributed teams of vehicles, sensors, and aircraft. In particular, my work investigates:

• Theories: Provably near-optimal tools for submodular optimization and online learning;
• Algorithms: Resource-aware coordination algorithms near-optimal performance and provably balances the trade-off of scalability and optimality in unstructured, untrustworthy, and even adversarial environments;
• Applications: Scalable real-world implementation of communication- and computation-intensive coordination tasks onboard resource-constrained hardware, such as active SLAM and target tracking using multiple UAVs.

Previously, I received an M.S. in Electrical & Computer Engineering from Georgia Tech in 2020 and a B.Eng. in Automation from Northeastern University (China) in 2018. I was also a Research Intern at Honda Research Institute in San Jose, CA for summer 2024.

email  /  google scholar  /  linkedin

• 03/10/2025: Received Rackham Predoctoral Fellowship to support the final year of my PhD studies!
• 07/24/2024: Paper Performance-Aware Self-Configurable Multi-Agent Networks: A Distributed Submodular Approach for Simultaneous Coordination and Network Design accepted to CDC 2024. [arxiv] [code] [slides] [presentation video]
• 05/20/2024: I start my first day in San Jose, CA this summer as a Research Intern at HRI. Feel free to drop by and hang out with me if you're around - I'd love to catch up and share what I've been working on!
• 01/22/2024: Paper on intelligently leveraging untrustworthy external commands for multi-robot coordination in unstructured environments accepted to ACC 2024.
• 08/31/2023: Paper Resource-Aware Simulator for Decentralized Active Perception with Multiple Robots accepted to IROS 2023 Workshop on Advances in Multi-Agent Learning - Coordination, Perception, and Control.

Journal Papers

1. Communication- and Computation-Efficient Distributed Submodular Optimization in Robot Mesh Networks
   Zirui Xu*, Sandilya Sai Garimella*, Vasileios Tzoumas
   under review, 2024.
   arxiv / code


2. Online and Robust Intermittent Motion Planning in Dynamic and Changing Environments
   Zirui Xu, George P. Kontoudis, Kyriakos G. Vamvoudakis
   IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2023.
   paper / simulation video


3. Online Submodular Coordination with Bounded Tracking Regret: Theory, Algorithm, and Applications to Multi-Robot Coordination
   Zirui Xu, Hongyu Zhou, Vasileios Tzoumas
   IEEE Robotics and Automation Letters (RA-L), 2023.
   arxiv / code

Conference Papers

1. Performance-Aware Self-Configurable Multi-Agent Networks: A Distributed Submodular Approach for Simultaneous Coordination and Network Design
   Zirui Xu, Vasileios Tzoumas
   IEEE Conference on Decision and Control (CDC), 2024. Invited Paper.
   arxiv / code / slides / presentation video


2. Leveraging Untrustworthy Commands for Multi-Robot Coordination in Unpredictable Environments: A Bandit Submodular Maximization Approach
   Zirui Xu*, Xiaofeng Lin*, Vasileios Tzoumas
   American Control Conference (ACC), 2024.
   arxiv


3. Resource-Aware Simulator for Decentralized Active Perception with Multiple Robots
   Sandilya Sai Garimella, Zirui Xu, Vasileios Tzoumas
   Workshop on Advances in Multi-Agent Learning - Coordination, Perception, and Control
   IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023.


4. Efficient Online Learning with Memory via Frank-Wolfe Optimization: Algorithms with Bounded Dynamic Regret and Applications to Control
   Hongyu Zhou, Zirui Xu, Vasileios Tzoumas
   IEEE Conference on Decision and Control (CDC), 2023.
   arxiv


5. Bandit Submodular Maximization for Multi-Robot Coordination in Unpredictable and Partially Observable Environments
   Zirui Xu, Xiaofeng Lin, Vasileios Tzoumas
   Robotics: Science and Systems (RSS), 2023.
   arxiv / code / simulation videos / presentation


6. Resource-Aware Distributed Submodular Maximization: A Paradigm for Multi-Robot Decision-Making
   Zirui Xu, Vasileios Tzoumas
   IEEE Conference on Decision and Control (CDC), 2022. Invited Paper.
   arxiv / code / presentation


7. Online, Model-Free Motion Planning in Dynamic Environments: An Intermittent, Finite Horizon Approach with Continuous-Time Q-Learning
   George P. Kontoudis, Zirui Xu, Kyriakos G. Vamvoudakis
   American Control Conference (ACC), 2020.
   pdf / simulation video / presentation

Book Chapters

1. RRT-QX: Real-Time Kinodynamic Motion Planning in Dynamic Environments with Continuous-Time Reinforcement Learning
   George P. Kontoudis, Kyriakos G. Vamvoudakis, Zirui Xu
   Brain and Cognitive Intelligence-Control in Robotics, ed. B. Wei, Chapter 1, CRC Press, 2022.
   book chapter / pdf / video

• Graduate Student Instructor: AEROSP 341 - Aircraft Dynamics, Winter 2024 & Winter 2025.
• Graduate Student Instructor: AEROSP 584 - Navigation and Guidance: From Perception to Control, Fall 2022.

• President of Graduate Student Advisory Committee (GSAC), Department of Aerospace Engineering, University of Michigan, 2023-24.
• Member of DEI Committee, Department of Aerospace Engineering, University of Michigan, 2022-23.

last updated: Mar 2025 | template from here