Junfeng (Eric) Chen (陈俊锋)

• Scalable distributed task planning for large-scale and highly dynamic system.
• Hybrid optimization and game-theoretic approaches for dynamic task planning under competition and cooperation.
• LLM-based and human-in-the-loop task planning for unknown and open-world scenarios.
• Communication-aware coordination for efficient and robust multi-robot collaboration.

Email / Google Scholar / Github

✉️ I'm expected to graduate in 2026 and open to both academia and industry positions. If you're interested, please feel free to contact me.

LLMs Meet Formal Methods for Robot Swarms: Reliable, Explainable and Efficient Human-in-the-loop Planning in Unknown Environments
Junfeng Chen, Yuxiao Zhu, An Zhuo, Xintong Zhang, Shuo Zhang, Meng Guo^#, and Zhongkui Li^#.
Submitted to Science Robotics (SR), 2025.

We propose a formal method and LLM framework for coordinating large fleets of heterogeneous robots in open and dynamic environments. Our approach integrates model-checking-based task planning with LLM-powered reasoning and interaction, ensuring adaptability, explainability, and optimal mission execution. Validated through simulations and real-world deployments, it proves effective for disaster response, infrastructure inspection, and dynamic surveillance.

AMBUSH: Collaborative Capture in Complex Environments with Neural Acceleration
Junfeng Chen, Yinhang Luo, Junrui Li, Xinyi Wang, and Meng Guo^#.
Submitted to IEEE Transactions on Robotics (T-RO), 2025.

We propose H-MCTS Ambush, a parameterized ambush strategy optimized with Hybrid Monte Carlo Tree Search for collaborative capture of dynamic targets. Our method accounts for topology, visibility, speed ratios, and capture ranges, while leveraging learned heuristics to accelerate planning without sacrificing quality, successfully capturing faster and even human-controlled evaders in complex environments.

CoCoPlan: Adaptive Coordination and Communication for Multi-robot Systems in Dynamic and Unknown Environments
Xintong Zhang*, Junfeng Chen*, Yuxiao Zhu, and Meng Guo^#.
Submitted to IEEE Robotics and Automation Letters (RA-L), 2025

We propose CoCoPlan, a unified framework that co-optimizes collaborative task planning and intermittent communication for multi-robot systems. Our approach integrates the branch-and-bound task encoding, adaptive efficiency objectives, and optimized event scheduling to handle dynamic environments under limited connectivity in both office and disaster-response scenarios.

SLEI3D: Simultaneous Exploration and Inspection via Heterogeneous Fleets under Limited Communication
Junfeng Chen, Yuxiao Zhu, Xintong Zhang, and Meng Guo^#.
Revision on IEEE Transactions on Automation Science and Engineering (T-ASE), 2025

We propose SLEI3D, a planning and coordination framework for heterogeneous multi-robot systems to perform simultaneous 3D exploration, inspection, and real-time reporting in unknown environments. Our approach integrates adaptive inspection and intermittent communication protocols with a multi-layer, multi-rate planning mechanism for robust coordination.

DEXTER-LLM: Dynamic and Explainable Coordination of Multi-Robot Systems in Unknown Environments via Large Language Models
Yuxiao Zhu*, Junfeng Chen*, Xintong Zhang, Meng Guo, and Zhongkui Li^#.
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025

We propose DEXTER-LLM, a novel framework for dynamic task planning in unknown environments. Our approach integrates LLM-based multi-stage reasoning, optimization-based task assignment, and adaptive human-in-the-loop verification to tackle the challenges of online adaptability and explainability.

Meta-reinforcement learning based cooperative surface inspection of 3d uncertain structures using multi-robot systems
Junfeng Chen, Yuan Gao, Junjie Hu, Fuqin Deng, and Tin Lun Lam^# .
IEEE International Conference on Robotics and Automation (ICRA), 2024.

We propose Meta-Scan, a meta-learning approach for efficient sensor scanning in 3D and uncertain environments using heterogeneous multi-robot systems. Our method enhances exploration and rapid adaptation in the complex scenarios.

Multirobolearn: An open-source framework for multi-robot deep reinforcement learning
Junfeng Chen, Fuqing Deng, Yuan Gao, Junjie Hu, Xiyue Guo, Guanqi Liang, and Tin Lun Lam^#.
IEEE International Conference on Robotics and Biomimetics (ROBIO), 2023.
Star

We propose MultiRoboLearn, an open-source framework that bridges multi-agent deep reinforcement learning with real-world multi-robot applications. Our framework offers standardized, easy-to-use simulation scenarios seamlessly transferable to physical multi-robot systems, and provides a benchmark for varying algorithm comparisons.

Accelerated K-Serial Stable Coalition for Dynamic Capture and Resource Defense
Junfeng Chen, Zili Tang, and Meng Guo^#.
IEEE Robotics and Automation Letters (RA-L), 2023.

We propose K-Serial Coalition, a distributed strategy for adaptive coalition formation in large-scale heterogeneous multi-robot systems. Our approach combines a K-serial stable coalition algorithm with a heterogeneous graph attention network heuristic to enhance online adaptability while guaranteeing solution quality.

Combinatorial-hybrid Optimization for Multi-agent Systems under Collaborative Tasks
Zili Tang, Junfeng Chen*, and Meng Guo^#.
IEEE Conference on Decision and Control (CDC), 2023.

We propose CHO-Framework, a combinatorial-hybrid optimization approach for multi-agent coordination. Our method jointly addresses coalition formation and collaborative control by interleaving discrete task assignment with continuous behavior optimization to guarantee the feasibility and quality.

Asymmetric self-play-enabled intelligent heterogeneous multirobot catching system using deep multiagent reinforcement learning
Yuan Gao, Junfeng Chen, Xi Chen, Chongyang Wang, Junjie Hu, Fuqin Deng, and Tin Lun Lam^#.
IEEE Transactions on Robotics (T-RO), 2023.

We propose ASP-CL Framework, an actor-critic multi-agent reinforcement learning approach that integrates asymmetric self-play and curriculum learning for heterogeneous multi-robot systems. Our method enables cooperative behaviors in adversarial catching tasks under real-world constraints.

Energy sharing mechanism for a freeform robotic system-freebot
Guanqi Liang, Yuxiao Tu, Lijun Zong, Junfeng Chen, and Tin Lun Lam^#.
IEEE International Conference on Robotics and Automation (ICRA), 2022.

We propose FreeBOT-Energy, an energy sharing mechanism for modular self-reconfigurable robots that enables modules to exchange power through surface contact. Our approach designs robust sharing rules and network structures to ensure sustainability and maximize participating modules.

AcousticFusion: Fusing sound source localization to visual SLAM in dynamic environments
Tianwei Zhang, Huayan Zhang, Xiaofei Li, Junfeng Chen, Tin Lun Lam, and Sethu Vijayakumar^#.
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021.

We propose AV-SLAM, an audio-visual fusion approach that integrates sound source direction with RGB-D sensing to handle dynamic obstacles in multi-robot SLAM. Our method replaces costly object detectors with lightweight DoA-based cues, enabling efficient on-board processing and achieving stable localization across diverse dynamic environments.

Semantic histogram based graph matching for real-time multi-robot global localization in large scale environment
Xiyue Guo, Junjie Hu, Junfeng Chen, Fuqin Deng, and Tin Lun Lam^#.
IEEE Robotics and Automation Letters (RA-L), 2021.

We propose SH-Graph, a semantic histogram-based graph matching method for efficient and robust multi-robot global localization. Our approach overcomes large viewpoint variations while enabling real-time performance in both homogeneous and heterogeneous robot systems.

A two-stage unsupervised approach for low light image enhancement
Junjie Hu, Xiyue Guo, Junfeng Chen, Guanqi, Liang, Fuqin Deng, and Tin Lun Lam^#.
IEEE Robotics and Automation Letters (RA-L), 2021.

We propose ULIE-Net, a two-stage unsupervised framework for low-light image enhancement. Our method combines Retinex-based pre-enhancement with an adversarially trained refinement network, overcoming the need for paired data, poor dark-scene performance, and noise amplification.

A novel mission planning method for UAVs’ course of action
Yaoming Zhou, Haoran Zhao, Junfeng Chen, and Yuhong Jia^#.
Computer Communication, 2020.

We propose NSG-Planner, a mission planning method for UAV courses of action based on a two-segment nested scheme generation strategy. Our approach integrates task decomposition and resource scheduling to automatically produce diverse operational schemes with superior task completion time to verify its effectiveness and flexibility.

[08/2024], Postgraduate Studentship.

[09/2018], First-class Graduate studentship.

[09/2017], First-class Graduate studentship.

[11/2015], National Scholarship - Ministry of Education, PRC.

[11/2014], National Scholarship - Ministry of Education, PRC.

Reviewer: T-RO, RA-L, T-ASE, JFR, IET Cyber-Systems and Robotics, ICRA, IROS

Teaching Assistant:

Skills: Python, C++, Matlab, C#, ROS, Machine Learning (ML), RL, Graph Learning, Git & GitHub, Robotics Simulator (Gazebo, Isaac Gym, Unity, V-rep, Mujoco, PyBullet), CATIA, AutoCAD, Robotic Platforms (Tello, Crazyflies, Quadrotor, TurtleBot, KKSwarms, Scout, AgileX, Go2), Heterogeneous Multi-Robot Systems.

A glimpse of our recent experimental activities and achievements

Last update: 2025.09.14. Thanks.