Junfeng Long

ShangHai JiAi Genetics & IVF Institute, University of California, Berkeley

Papers

Total Citations

H-Index

About

Junfeng Long is a rising star in humanoid robotics, whose work tackles the fundamental challenge of making bipedal machines agile and stable in the real world. His primary research focuses on perceptive locomotion control and accessible hardware design. Long’s most impactful contribution is the **Perceptive Internal Model**, a framework that integrates visual perception directly into the control loop, enabling humanoid robots to navigate complex terrains with unprecedented stability—a critical advance over “blind” quadruped policies. His **Hybrid Internal Model** further refines this by fusing simulated robot responses with noisy sensor data to improve state estimation for agile locomotion. Beyond algorithms, Long is a champion of open science: he led the development of the **Berkeley Humanoid Lite**, an open-source, 3D-printed humanoid robot that dramatically lowers the cost barrier to entry in the field. This work, already cited over a dozen times in just two years, demonstrates his dual impact—pushing both the theoretical frontiers of control and the practical democratization of hardware. For students and researchers, Long exemplifies how combining elegant control theory with accessible engineering can accelerate the entire field.