Project 24: Learning-based Multi-robot Cooperation Design and Analysis

Contact Information:

Assoc. Prof. Jianping He

Project Description and Objectives:

In recent years, there has been a growing interest in multi-robot cooperative control systems, thanks to their remarkable capabilities in handling complex tasks. However, the efficiency and safety issues associated with multi-robot systems are significant, due to their highly distributed nature and openness to the unknown external environment in real-world scenarios.

In this research project, you will get a basic understanding of the multi-robot systems and learn how to develop an efficient strategy to improve the cooperation performance through state-of-the-art techniques, e.g., reinforcement learning and model-free optimization. Additionally, you will explore the study on the performance analysis to reveal the strength of your proposed designs in a systematic approach. In the end, you can validate the effectiveness of the proposed methods by testing the algorithm in both simulation and real-world robot platforms.

Eligibility Requirements:

Interested students should be proficient in Python/C++ programming and have basic knowledge of robot control. In addition, proficiency in English writing and speaking is mandatory.

Main Tasks:

Propose a new idea for the learning-based cooperation of multi-robot system.

Present the result with either a simulation or an experiment.

Finish a report of the internship.

Website:

Lab: http://iwin-fins.com

School: http://www.seiee.sjtu.edu.cn/