Avatar

Mohsen Sombolestan

Control and Robotics Lead

Proception AI

I am the Control and Robotics Lead at Proception AI, where I lead the development of control and robotics software for a bimanual dexterous manipulation robot, spanning real-time control, teleoperation, and the data pipeline behind learned manipulation policies.

I completed my PhD in Robotics at the University of Southern California (USC), advised by Prof. Quan Nguyen. My research focused on advancing legged locomotion and manipulation in uncertain environments, leveraging model predictive control (MPC) and adaptive control techniques.

I was previously a research intern at Field AI, where I worked on addressing vision-based locomotion challenges for quadruped robots navigating complex terrains, such as those found on construction sites.

Education

  • University of Southern California

    PhD in Mechanical Engineering, Robotics and Control

  • Isfahan University of Technology

    MSc in Mechanical Engineering, Robotics and Control

  • Sharif University of Technology

    BSc in Mechanical Engineering

News

Publications

Hierarchical Adaptive Motion Planning with Nonlinear Model Predictive Control for Safety-Critical Collaborative Loco-Manipulation

Preprint, 2024, submitted to The International Journal of Robotics Research (IJRR)

Adaptive Force-Based Control of Dynamic Legged Locomotion over Uneven Terrain

IEEE Transactions on Robotics (T-RO), 2024

Hierarchical Adaptive Control for Collaborative Manipulation of a Rigid Object by Quadrupedal Robots

IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, 2023

Hierarchical Adaptive Loco-manipulation Control for Quadruped Robots

IEEE/RSJ International Conference on Robotics and Automation, ICRA, 2023

Adaptive Force-based Control for Legged Robots

IEEE/RSJ International Conference on Robotics and Automation, IROS, 2021

Optimal path-planning for mobile robots to find a hidden target in an unknown environment based on machine learning

Journal of Ambient Intelligence and Humanized Computing, 2018

Ongoing Projects

Adaptive Loco-Manipulation Control for Humanoid Robots

Expanding our methodology to empower humanoid robots (such as HECTOR) with capabilities for performing loco-manipulation tasks and perceptive locomotion across diverse and challenging terrains.

Adaptive Sampling-based Model Predictive Control for Bipedal Locomotion

Exploring the potential of adaptive sampling-based MPC for bipedal robots to address the challenges of real-time planning and control in short-horizon agile movements and disturbance rejection under model uncertainty.