Songjun Huang

About

Songjun Huang is a rising researcher at the forefront of intelligent multi-robot systems, specializing in reinforcement learning (RL), task allocation, and path planning under uncertainty. His work addresses critical challenges in autonomous decision-making for complex, real-world missions, particularly in post-disaster rescue and informed search scenarios. Huang’s major contributions include pioneering multi-behavior multi-agent RL frameworks that enable robots to learn coordinated strategies from offline training data, significantly improving search efficiency in unknown environments. He has also developed bi-layer joint training architectures that integrate task allocation and path-finding, achieving robust performance despite environmental unpredictability. With over 20 citations across his most-cited papers from 2024-2025, his research is gaining rapid traction. Notably, his recent work on Retrieval-Augmented Hierarchical in-context RL (RAHL) bridges large language models with hierarchical reinforcement learning, enabling robots to leverage LLM-based reasoning for complex task planning. This innovative fusion of language models and RL positions Huang as a key contributor to the next generation of adaptive, intelligent robotic systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Multi-Behavior Multi-Agent Reinforcement Learning for Informed Search via Offline Training

   10 citations · 2024
2. 2
   Reinforcement learning–based task allocation and path‐finding in multi‐robot systems under environment uncertainty

   4 citations · 2025
3. 3
   A Bi-Layer Joint Training Reinforcement Learning Framework for Post-Disaster Rescue

   4 citations · 2024
4. 4
   Retrieval-Augmented Hierarchical in-Context Reinforcement Learning and Hindsight Modular Reflections for Task Planning with LLMs

   2 citations · 2025

Key Collaborators

Contact & Links