Songyuan Zhang
— After all this time? Always.
I am a Ph.D. student at REALM lab led by Prof. Chuchu Fan in Department of Aeronautics and Astronautics, MIT. I received my Master of Science in Aeronautics and Astronautics at MIT in 2024, and Bachelor of Engineering at Tsien Excellence in Engineering Program (Tsien class, TEEP) at Tsinghua University in 2021. The goal of my research is to develop real intelligent safe assistive robots. My research interests lie in reinforcement learning (offline/online), safe control for robotics, neural certificates, generative AI, multi-agent systems, control theories, and robotics. I interned with the Amazon Industrial Robotics Group in summer 2026. I am an Amazon AI PhD Fellow, and a recipient of the RSS Outstanding Student Paper Award (2025).
Besides research, I enjoy traveling, hiking, photography, skiing, badminton, and running. I am also a Gryffindor!
News
- Jul 15, 2026: Our paper “Handbook on Safety Certificates: Techniques from Hamilton-Jacobi Reachability Analysis and Control Barrier Functions” is accepted by the 2026 65th IEEE Conference on Decision and Control (CDC)!
- Jul 11, 2026: Our paper Safe and Scalable Multi-Drone Payload Transport via CBF-based Reinforcement Learning with Zero-Shot Sim-to-Real Transfer is accepted by IEEE Robotics and Automation Letters (RA-L)!
- Jun 23, 2026: Our paper Hamiltonian Grid-Forming: Methodology and Applications for IBR-Rich Grid Operations is accepted by IEEE Transactions on Industry Applications!
- Jan 31, 2025: Our paper “Beyond Waypoints: Semantic-Centric Autonomy with Unreliable Maps Through Learned Abstractions” is accepted by ICRA 2026!
- Jan 26, 2025: Our papers ReFORM: Reflected Flows for On-support Offline RL via Noise Manipulation and Solving Parameter-Robust Avoid Problems with Unknown Feasibility using Reinforcement Learning are accepted by ICLR 2026!
- Sep 18, 2025: Our paper HMARL-CBF – Hierarchical Multi-Agent Reinforcement Learning with Control Barrier Functions for Safety-Critical Autonomous Systems is accepted by NeurIPS 2025!