Hello there! I am a visiting scholar at Hybrid Robotics, UC Berkeley advised by Prof. Koushil Sreenath. Previously, I worked on Legged Robot Locomotion with Dr. Jiangmiao Pang and Distributed Systems with Prof. Youlong Wu. I received my Bachelor's degree in Computer Science from ShanghaiTech University.

I am interested in enabling robotic systems to achieve functionality and safety. Toward this goal, I focus on: (a) bridging control and machine learning to develop methods possessing their advantages; (b) training and application of dynamics models (world models); (c) developing robot foundation model for general purpose robots.

I am actively applying for Ph.D. position starting from Fall 2025. If you are interested in my research or want to chat, please drop me an email.

Google Scholar  |  Github  |  Twitter  |  Email 

Junfeng Long「龙俊峰」

We proposed GRUtopia, the first simulated interactive 3D society designed for various robots. It features (a)GRScenes, a dataset with 100k interactive and finely annotated scenes. (b) GRResidents, a LLM driven NPC system. (c) GRBench, a benchmark posing moderately challenging tasks.

We propose the Perceptive Intenal Model (PIM), a method to estimate environmental disturbances with perceptive information, enabling agile and robust locomotion for various humanoid robots on various terrains.

We propose the Parallelized Model-based Reinforcement Learning (PaMoRL) framework. PaMoRL introduces two novel techniques: the Parallel World Model (PWM) and the Parallelized Eligibility Trace Estimation (PETE) to parallelize both model and policy learning stages of current MBRL methods over the sequence length.

We present the H-infinity Locomotion Control, an adversarial framework improving the control policy's ability to resist external disturbances with H-infinity performance guarantee.

We propose TOP-Nav, a novel legged navigation framework that integrates a comprehensive path planner with Terrain awareness, Obstacle avoidance and close-loop Proprioception.

We present the Hybrid Internal Model, a method enabling the control policy to estimate environmental disturbances by only explicitly estimating velocity and implicitly simulating the system's response.

We propose a coded scheme to reduce the communication latency by exploiting computation and communication capabilities of all nodes and creating coded multicast opportunities.More importantly, we prove that the proposed scheme is always optimal, i.e., achieving the minimum communication latency, for arbitrary computing and storage abilities at the master.

Updated at November 2024.
Template from Jon Barron.