Quentin Le Lidec

About

Quentin Le Lidec is a robotics researcher whose work sits at the intersection of physics simulation, differentiable programming, and robot control. His research focuses on two deeply interconnected challenges: making physics simulators more accurate and differentiable, and leveraging these capabilities to improve robot learning and optimization. Le Lidec has made significant contributions to differentiable simulation, particularly for systems involving frictional contacts — notoriously difficult to model and differentiate. His 2021 paper on physical system identification (44 citations) demonstrated how differentiable physics can enable more principled model-based reinforcement learning and optimal control. Complementing this, his work on randomized smoothing offers elegant solutions for handling non-differentiable physical processes, making gradient-based optimization more robust in practice. A distinctive thread in his research is collision detection, where he has reimagined the classical GJK algorithm through an optimization lens. His accelerated variants and differentiable formulations (collectively accumulating over 70 citations) have meaningfully advanced both speed and gradient accessibility in this foundational problem. His comprehensive comparative analysis of contact models in robotics further cements his role as a synthesizer and evaluator of the field's landscape. Across fewer than five years of output, Le Lidec has established himself as a thoughtful contributor bridging the gap between physically accurate simulation and data-driven robot control.

Research Focus

Key Achievements

Top Papers

1. 1
   Differentiable Simulation for Physical System Identification

   44 citations · 2021
2. 2
   Contact Models in Robotics: A Comparative Analysis

   36 citations · 2024
3. 3
   GJK++: Leveraging Acceleration Methods for Faster Collision Detection

   27 citations · 2024
4. 4
   Differentiable Collision Detection: a Randomized Smoothing Approach

   26 citations · 2023
5. 5
   Collision Detection Accelerated: An Optimization Perspective

   20 citations · 2022
6. 6
   From Compliant to Rigid Contact Simulation: a Unified and Efficient Approach

   9 citations · 2024
7. 7
   Contact Models in Robotics: a Comparative Analysis

   4 citations · 2023
8. 8
   Augmenting differentiable physics with randomized smoothing

   3 citations · 2022
9. 9
   Differentiable Simulation of Soft Robots with Frictional Contacts

   2 citations · 2025
10. 10
    Enforcing the consensus between Trajectory Optimization and Policy Learning for precise robot control

    2 citations · 2023

Key Collaborators

Contact & Links