Embedding Physics Priors in Robot Learning: A Survey
cs.RO
Submitted: 2026-09-15
Updated: 2026-09-15
Code: https://github.com/TUM-AVS/survey-physics-embeddedrobot-learning
Terminology
Sources
- Model-Based Control Using Koopman Operators
- Learning-Based Modeling of Soft Robots via Cosserat Rod Theory
- Structure-Preserving Learning of Nonholonomic Dynamics
- Language Models are Few-Shot Learners
- Neural Ordinary Differential Equations
- Learning Robotic Manipulation Policies from Point Clouds with Conditional Flow Matching
- Lagrangian Neural Networks
- Physically Native World Models: A Hamiltonian Perspective on Generative World Modeling
- Physically Consistent Modeling & Identification of Nonlinear Friction with Dissipative Gaussian Processes
- Neural Networks with Physics-Informed Architectures and Constraints for Dynamical Systems Modeling
- How to Learn and Generalize From Three Minutes of Data: Physics-Constrained and Uncertainty-Aware Neural Stochastic Differential Equations
- Adaptive Control of SE(3) Hamiltonian Dynamics with Learned Disturbance Features
- Simplifying Hamiltonian and Lagrangian Neural Networks via Explicit Constraints
- Brax -- A Differentiable Physics Engine for Large Scale Rigid Body Simulation
- Hamiltonian Neural Networks
- A General Framework for Structured Learning of Mechanical Systems
- Forced Variational Integrator Networks for Prediction and Control of Mechanical Systems
- NeuralSim: Augmenting Differentiable Simulators with Neural Networks
- NVIDIA SimNet^{TM}: an AI-accelerated multi-physics simulation framework
- Deeptime: a Python library for machine learning dynamical models from time series data
Related papers
- FMT x: An Efficient and Asymptotically Optimal Extension of the Fast Marching Tree for Dynamic Replanning
- MPCFormer: A physics-informed data-driven approach for explainable socially-aware autonomous driving
- RoboLab: A High-Fidelity Simulation Benchmark for Analysis of Task Generalist Policies
- HRDexDB: A 4D Dexterous Grasping Dataset Across Human and Multiple Robot Embodiments
- APT: Action Expert Pretraining Improves Instruction Generalization of Vision-Language-Action Policies
- Fine-tuning is Not Enough: A Parallel Framework for Collaborative Imitation and Reinforcement Learning in End-to-end Autonomous Driving