Contact Information


Research Interests

  • Machine Learning
  • Multi-agents coordination
  • Reinforcement Learning

Haoxuan Pan

Class of 2021

Shanghai Jiao Tong University

I am an M.S. student at the Department of Automation, Shanghai Jiao Tong University (SJTU), supervised by Prof. Jianping He and Prof. Xiaoming Duan. I received the honored B.Eng. degree in Automation from XJTU in 2021. I was selected for the 2022 Tencent Rhino-Bird Research Elite Program (only 70+ students were selected). I worked with Dr. Deheng Ye and Prof. Mingfei Sun during my internship at Tencent.

My research interests lie in Reinforcement Learning and Multi-agents coordination.


Revisiting Estimation Bias in Policy Gradients for DRL

We revisit the important issue of estimation bias of PG for DRL. Check our papers at

This research topic was done with the help of Tencent AI Lab and Dr. Mingfei Sun at the University of Manchester.

Real-time digital twin platform of multi-robots

We are trying to research the sim-to-real problem of robots in Reinforcement Learning. We now finished the first step target–the robot digital twin robot platform.

The digital twin robot platform

We constructed a real-time digital twin platform to explore the multi-agent coordination problem and how the virtual environment can help refine the real environment’s outcomes. The following picture is an overview of the platform:

Overview of the Robot Platform

I am mainly responsible for the information exchange parts, including multiple data fusion, multi-thread concurrency, backend, and database construction, and the visualization part. The following image shows the details and my contributions to the platform. The green parts are mostly designed and implemented by myself, and the yellow parts are those completed by me together with other participants.

Details of the Robot Platform

Awards and Honors

  • 2021: Outstanding Graduate of Xi’an Jiaotong University

Social Media