Simulation Fidelity

Simulation fidelity, the accuracy of a simulation in representing a real-world system, is crucial for various applications, particularly in robotics and AI, where it enables training and testing of algorithms in a controlled environment before real-world deployment. Current research focuses on optimizing the balance between simulation fidelity and computational cost, exploring techniques like multi-fidelity approaches that leverage both high-fidelity and low-fidelity models, and developing more efficient training methods for deep learning surrogates of complex physical systems. Improved simulation fidelity leads to more robust and reliable AI systems, impacting fields ranging from autonomous vehicle development to biomedical applications like targeted drug delivery.

Papers

July 11, 2024

A Review of Nine Physics Engines for Reinforcement Learning Research
Michael Kaup, Cornelius Wolff, Hyerim Hwang, Julius Mayer, Elia Bruni
Reinforcement Learning Simulation Environment Reinforcement Learning Research Simulation Platform Physic Engine Simulation Fidelity

May 21, 2024

Better Simulations for Validating Causal Discovery with the DAG-Adaptation of the Onion Method
Bryan Andrews, Erich Kummerfeld
Causal Discovery Causal Model Causal Structure Learning Simulation Fidelity Dag DLT

March 29, 2024

Path planning of magnetic microswimmers in high-fidelity simulations of capillaries with deep reinforcement learning
Lucas Amoudruz, Sergey Litvinov, Petros Koumoutsakos
Deep Reinforcement Learning Path Planning Biomedical Application Simulation Fidelity Microvascular Network

February 22, 2024

Autonomy Oriented Digital Twins for Real2Sim2Real Autoware Deployment
Chinmay Vilas Samak, Tanmay Vilas Samak
Autonomous Vehicle Digital Twin Legal Autonomy Simulation Fidelity

September 28, 2023

High Throughput Training of Deep Surrogates from Large Ensemble Runs
Lucas Meyer, Marc Schouler, Robert Alexander Caulk, Alejandro Ribés, Bruno Raffin
Deep Learning Numerical Solver Reservoir Dynamic Training Throughput Deep Surrogate Simulation Fidelity Simulation Ensemble

December 28, 2022

Falsification of Learning-Based Controllers through Multi-Fidelity Bayesian Optimization
Zahra Shahrooei, Mykel J. Kochenderfer, Ali Baheri
Learning Based Multi Fidelity Bayesian Optimization Simulation Fidelity

November 22, 2022

A Reinforcement Learning Badminton Environment for Simulating Player Tactics (Student Abstract)
Li-Chun Huang, Nai-Zen Hseuh, Yen-Che Chien, Wei-Yao Wang, Kuang-Da Wang, Wen-Chih Peng
Player Performance Turn Based Simulation Fidelity

November 21, 2022

Exploring Physical Latent Spaces for High-Resolution Flow Restoration
Chloe Paliard, Nils Thuerey, Kiwon Um
Latent Space Sufficient Representation Flow Reconstruction Simulation Fidelity

July 26, 2022

Semi-analytical Industrial Cooling System Model for Reinforcement Learning
Yuri Chervonyi, Praneet Dutta, Piotr Trochim, Octavian Voicu, Cosmin Paduraru, Crystal Qian, Emre Karagozler, Jared Quincy Davis, Richard Chippendale, Gautam Bajaj, Sims Witherspoon, Jerry Luo
Reinforcement Learning Reinforcement Learning Algorithm Reinforcement Learning Research Simulation Fidelity Multi Physic Simulation Industrial Cooling System

July 21, 2022

Rethinking Sim2Real: Lower Fidelity Simulation Leads to Higher Sim2Real Transfer in Navigation
Joanne Truong, Max Rudolph, Naoki Yokoyama, Sonia Chernova, Dhruv Batra, Akshara Rai
Simulation Study Local Navigation Robot Motion Fast Simulation Fidelity Model Sim2Real Transfer Simulation Fidelity Sim2Real Gap

Simulation Fidelity

Papers

A Review of Nine Physics Engines for Reinforcement Learning Research

Better Simulations for Validating Causal Discovery with the DAG-Adaptation of the Onion Method

Path planning of magnetic microswimmers in high-fidelity simulations of capillaries with deep reinforcement learning

Autonomy Oriented Digital Twins for Real2Sim2Real Autoware Deployment

High Throughput Training of Deep Surrogates from Large Ensemble Runs

Falsification of Learning-Based Controllers through Multi-Fidelity Bayesian Optimization

A Reinforcement Learning Badminton Environment for Simulating Player Tactics (Student Abstract)

Exploring Physical Latent Spaces for High-Resolution Flow Restoration

Semi-analytical Industrial Cooling System Model for Reinforcement Learning

Rethinking Sim2Real: Lower Fidelity Simulation Leads to Higher Sim2Real Transfer in Navigation