EXO-PLORE: Exploring Exoskeleton Control Space through Human-Aligned Simulation, ICLR 2026

Official implementation of Trust Region Inverse RL (TRIRL) in Jax

MPC-guided RL for Humanoid Locomotion and Manipulation

[CVPR 2024 Highlight] FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects

A Proprioceptive-Only Benchmark for Quadruped Robot's State Estimation

Sumo: Dynamic and Generalizable Whole-Body Loco-Manipulation

PAL Robotics' robots in mjlab

Official code release for ICLR26 "XQC: Well-conditioned Optimization Accelerates Deep Reinforcement Learning"

Open-source code of Floating-Base Deep Lagrangian Networks (FeLaN).

High-Level Control Library for Franka Robots with Python and C++ Support

a legged robot (Unitree Go1) dataset containing leg kinematics (joint encoders and contact sesors), imu and lidar

🪶iterated Shared Soft Actor-Critic [ICLR 26], a new algorithm improving the sample-efficiency of target-free algorithms (e.g. SimbaV2) to bridge the gap with target-based algorithms🪶

Learning Legged MPC with Smooth Neural Surrogates

Algorithms for implementation the Explicit MPC using the multi-parametric quadratic programming.

A flexible optimization framework for whole-body loco-manipulation, built with Pinocchio and CasADi. Supports multiple dynamics formulations and solver backends.

Cross-State Transition Attention Transformer for improved robotic manipulation with better temporal modeling; https://arxiv.org/abs/2510.00726

Context-Aware Deep Lagrangian Networks for Model Predictive Control (CaDeLaC)

An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.

Release repo for our SLAM Handbook

A model predictive controller for quadruped robots based on the single rigid body model and written in python. Gradient-based (acados) or Sampling-based (jax)

Repository for the "Effective Design of Graphics for (Robotics) Research" tutorial

A Python package for linear subspace identification, nonlinear system identification, and nonlinear regression using Jax

Model Predictive Control in JAX

Official Implementation of "Sampling-Based System Identification with Active Exploration for Legged Robot Sim2Real Learning"

A robust algorithm for point cloud registration, using Graduated Non-Convexity and a Geman-McClure robust loss function to handle high outlier rates.

Must-read papers on Reinforcement Learning (RL)

A State Estimation Package for Quadruped Robots, that fuses Proprioceptive and Exteroceptive data