Yipeng Wang

profile_pic.jpg

Horseshoe Bend, AZ


I am an incoming MSE Robotics student at UPenn GRASP Lab. I recently graduated from NYU Tandon with a BS in Electrical and Computer Engineering.


My research interests lie at the intersection of robot learning and spatial intelligence for embodied agents. During my time at NYU, I am fortunate to be advised by Prof. Chen Feng at AI4CE Lab on non-prehensile mobile manipulation and sim-to-real robot learning. My work has focused on building robotic systems that connect simulation, perception, and control, with an emphasis on making learned policies deployable on hardware.

Publications

  1. RL+IL, Sim2Real
    egopush_v1_x4.gif
    EgoPush: Learning End-to-End Egocentric Multi-Object Rearrangement for Mobile Robots
    Boyuan An, Zhexiong Wang*, Yipeng Wang*, Jiaqi Li, Sihang Li†, Jing Zhang†, and Chen Feng†
    arXiv preprint arXiv:2602.18071, 2026
    arXiv Website

Projects

  • Quadrupedal Locomotion for Velocity Tracking
    Quadrupedal Locomotion for Velocity Tracking
    Learned quadrupedal locomotion policy with gait shaped by Raibert Heuristics with PPO in IsaacLab for Unitree Go2 that tracks velocity commands under joint disturbances produced by a custom actuator friction model
    ROB-GY 6323 Reinforcement Learning and Optimal Control for Autonomous Systems I, 2025
  • Omnidirectional Quadrupedal Landing Control
    Omnidirectional Quadrupedal Landing Control
    Learned aerial reorientation policy with PPO in MuJoCo Playground that lands Unitree Go1 from any orientation by redistributing rotational inertia under conserved angular momentum
    ROB-GY 7863 Planning, Learning, and Control for Space Robotics, 2025
  • SQP Solver for 2D Quadrotor Aerobatics
    SQP Solver for 2D Quadrotor Aerobatics
    A From-scratch SQP trajectory optimizer for 2D quadrotor aerobatics, extended into a MPC that robustly executes triple-flip maneuvers under rotor disturbances
    ROB-GY 6323 Reinforcement Learning and Optimal Control for Autonomous Systems I, 2025