avatar

Victor Dhédin

PhD Candidate
Technical University of Munich
victor.dhedin@tum.de


About Me

I’m a PhD candidate at the Technical University of Munich (TUM), advised by Prof. Majid Khadiv, with a background in machine learning and robotics. My research focuses on solving loco-manipulation tasks with a humanoid robot.

Before that, I received my Master’s degree in Data and Information Sciences from CentraleSupélec, and spent time at the Max Planck Institute for Intelligent Systems, the French Atomic Energy Commission (CEA) and Thales.

I'm exploring a potential transition from academia to industry. Feel free to reach out.

Research Interests

News

Education

Publications

* equal contribution † project lead ‡ corresponding author

|
  1. Humanoid-DART: Humanoid Loco-Manipulation using Diffusion-guided Augmentation through Relabeling and Tracking
    Pranav Debbad*, Kanish Thiagarajan*, Victor Dhédin*, Shafeef Omar, Majid Khadiv
    IEEE Robotics and Automation Letters (RA-L), 2026.
    arXiv PDF Video
  2. DynaRetarget: Dynamically-Feasible Retargeting using Sampling-Based Trajectory Optimization
    Victor Dhédin*, Ilyass Taouil*, Shafeef Omar*, Dian Yu, Kun Tao, Angela Dai, Majid Khadiv
    IEEE Robotics and Automation Letters (RA-L), 2026.
    IEEE arXiv PDF Code Project Page Video
  3. Simultaneous Contact Sequence and Patch Planning for Dynamic Locomotion
    Victor Dhédin, Haizhou Zhao, Majid Khadiv
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2025.
    arXiv PDF Video
  4. Diffusion-based Learning of Contact Plans for Agile Locomotion
    Victor Dhédin, Adithya Kumar Chinnakkonda Ravi, Armand Jordana, Huaijiang Zhu, Avadesh Meduri, Ludovic Righetti, Bernhard Schölkopf, Majid Khadiv
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2024.
    arXiv PDF Video
  5. Visual-Inertial and Leg Odometry Fusion for Dynamic Locomotion
    Victor Dhédin, Haolong Li, Shahram Khorshidi, Lukas Mack, Adithya Kumar Chinnakkonda Ravi, Avadesh Meduri, Paarth Shah, Felix Grimminger, Ludovic Righetti, Majid Khadiv, Jörg Stückler
    IEEE International Conference on Robotics and Automation (ICRA), 2023.
    arXiv PDF Video
  1. Humanoid-DART: Humanoid Loco-Manipulation using Diffusion-guided Augmentation through Relabeling and Tracking
    Pranav Debbad*, Kanish Thiagarajan*, Victor Dhédin*, Shafeef Omar, Majid Khadiv
    IEEE Robotics and Automation Letters (RA-L), 2026.
    arXiv PDF Video
  2. DynaRetarget: Dynamically-Feasible Retargeting using Sampling-Based Trajectory Optimization
    Victor Dhédin*, Ilyass Taouil*, Shafeef Omar*, Dian Yu, Kun Tao, Angela Dai, Majid Khadiv
    IEEE Robotics and Automation Letters (RA-L), 2026.
    IEEE arXiv PDF Code Project Page Video
  3. Simultaneous Contact Sequence and Patch Planning for Dynamic Locomotion
    Victor Dhédin, Haizhou Zhao, Majid Khadiv
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2025.
    arXiv PDF Video
  4. Task and Motion Planning for Humanoid Loco-manipulation teaser
    Task and Motion Planning for Humanoid Loco-manipulation
    Michal Ciebielski, Victor Dhédin, Majid Khadiv
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2025.
    arXiv PDF Video
  5. Diffusion-based Learning of Contact Plans for Agile Locomotion
    Victor Dhédin, Adithya Kumar Chinnakkonda Ravi, Armand Jordana, Huaijiang Zhu, Avadesh Meduri, Ludovic Righetti, Bernhard Schölkopf, Majid Khadiv
    IEEE-RAS International Conference on Humanoid Robots (Humanoids), 2024.
    arXiv PDF Video
  6. Learning Feasible Transitions for Efficient Contact Planning teaser
    Learning Feasible Transitions for Efficient Contact Planning
    Rikhat Akizhanov, Victor Dhédin, Majid Khadiv, Ivan Laptev
    Learning for Dynamics and Control Conference (L4DC), 2024.
    arXiv PDF
  7. Visual-Inertial and Leg Odometry Fusion for Dynamic Locomotion
    Victor Dhédin, Haolong Li, Shahram Khorshidi, Lukas Mack, Adithya Kumar Chinnakkonda Ravi, Avadesh Meduri, Paarth Shah, Felix Grimminger, Ludovic Righetti, Majid Khadiv, Jörg Stückler
    IEEE International Conference on Robotics and Automation (ICRA), 2023.
    arXiv PDF Video

Software

  1. SBTO: Sampling-Based Trajectory Optimization GitHub stars
    Official implementation of the sampling-based trajectory optimization behind DynaRetarget. It refines imperfect kinematic trajectories into dynamically feasible humanoid motions, incrementally extending the optimization horizon for long-horizon tasks.
    Language: Python, MuJoCo
    Code Project Page
  2. ATARI_NMPC GitHub stars
    Modular Nonlinear Model Predictive Control (NMPC) framework for quadruped robots, built on the Acados solver and Pinocchio. Easily adapted to new robots through URDF, gait and controller configuration.
    Language: Python, Acados, Pinocchio
    Code

Experience


Academic Experience

  1. Technical University of Munich
    PhD Candidate: 2023 – Present.
    Topics: Humanoid Loco-Manipulation, Synthetic Data Generation, Trajectory Optimization, Contact Planning.
    Advisor: Prof. Majid Khadiv.
  2. Max Planck Institute for Intelligent Systems, Tübingen
    Research Assistant: Jan. – Jun. 2022.
    Integrated a Visual-Inertial Odometry (VIO) algorithm on a Solo12 quadruped robot, with extensive real-world indoor and outdoor experiments. Published at ICRA 2023.

Industry Experience

  1. French Atomic Energy Commission (CEA), Pessac
    Research Intern: May – Nov. 2023.
    Master Thesis: Parametrization of robotic actions through demonstrations (graded A+, 20/20).
    Developed a graphical interface for novice users to program robotic arms for assembly tasks via kinesthetic demonstrations, with real-robot experiments and a user study. Led to a publication in Fusion Engineering and Design 2025.
  2. Thales, Palaiseau
    Research Engineer Intern: Jun. – Dec. 2021.
    Developed a proof-of-concept deep learning method for reconstructing corrupted audio recordings. Recorded and processed terabytes of audio data from an experimental optical-fiber distributed acoustic sensor (DAS).

Academic Service

Teaching

Reviewer


Powered by Jekyll and Minimal Light theme.