Peifeng Jiang

I am a Master’s student at Peking University, driven by a passion for building the next generation of robots that can intelligently perceive and interact with the physical world. My research is centered at the intersection of SLAM, Embodied AI, and 3D Gaussian Splatting, where I focus on closing the loop from perception to action.

My academic journey began with a deep dive into system optimization. As an undergraduate, I architected a high-performance parallel computing framework that boosted the performance of mainstream SLAM algorithms by 250%. Now, as a graduate student, my focus has shifted to leveraging the cutting-edge potential of 3DGS. I led the development of µSLAM, a robust, real-time (>30 FPS) dense SLAM system. Through its novel uncertainty-driven framework, µSLAM maintains state-of-the-art accuracy and stability even in challenging real-world environments. I am also pioneering a new 3D scene representation called the “Gaussian Memory Field,” aiming to equip robots with an interactive and interpretable world model.

For me, the excitement of research lies not just in theoretical innovation but in translating algorithms into robust, real-world systems. My engineering experience spans building robotic systems from scratch, bridging the gap between simulation (Unreal Engine) and real-world deployment (DashGo robots), and designing distributed perception algorithms for multi-robot clusters. I am passionate about tackling tough challenges, whether it’s addressing motion blur in humanoid SLAM or developing visual-inertial navigation for an unmanned supermarket.

My goal is to advance embodied AI by enabling robots to not only see the world but to truly understand and act within it. I believe a powerful, differentiable 3D world model is the key to unlocking this future. I am always eager to connect with fellow researchers and explore opportunities where these technologies can solve practical challenges. Feel free to reach out!