Deep Dynamics: Vehicle Dynamics Modeling with a Physics-Constrained Neural Network for Autonomous Racing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chrosniak, John, Ning, Jingyun, Behl, Madhur
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910723211264000
author Chrosniak, John
Ning, Jingyun
Behl, Madhur
author_facet Chrosniak, John
Ning, Jingyun
Behl, Madhur
contents Autonomous racing is a critical research area for autonomous driving, presenting significant challenges in vehicle dynamics modeling, such as balancing model precision and computational efficiency at high speeds (>280km/h), where minor errors in modeling have severe consequences. Existing physics-based models for vehicle dynamics require elaborate testing setups and tuning, which are hard to implement, time-intensive, and cost-prohibitive. Conversely, purely data-driven approaches do not generalize well and cannot adequately ensure physical constraints on predictions. This paper introduces Deep Dynamics, a physics-constrained neural network (PCNN) for vehicle dynamics modeling of an autonomous racecar. It combines physics coefficient estimation and dynamical equations to accurately predict vehicle states at high speeds and includes a unique Physics Guard layer to ensure internal coefficient estimates remain within their nominal physical ranges. Open-loop and closed-loop performance assessments, using a physics-based simulator and full-scale autonomous Indy racecar data, highlight Deep Dynamics as a promising approach for modeling racecar vehicle dynamics.
format Preprint
id arxiv_https___arxiv_org_abs_2312_04374
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Deep Dynamics: Vehicle Dynamics Modeling with a Physics-Constrained Neural Network for Autonomous Racing
Chrosniak, John
Ning, Jingyun
Behl, Madhur
Robotics
Artificial Intelligence
Machine Learning
I.2.9; I.6
Autonomous racing is a critical research area for autonomous driving, presenting significant challenges in vehicle dynamics modeling, such as balancing model precision and computational efficiency at high speeds (>280km/h), where minor errors in modeling have severe consequences. Existing physics-based models for vehicle dynamics require elaborate testing setups and tuning, which are hard to implement, time-intensive, and cost-prohibitive. Conversely, purely data-driven approaches do not generalize well and cannot adequately ensure physical constraints on predictions. This paper introduces Deep Dynamics, a physics-constrained neural network (PCNN) for vehicle dynamics modeling of an autonomous racecar. It combines physics coefficient estimation and dynamical equations to accurately predict vehicle states at high speeds and includes a unique Physics Guard layer to ensure internal coefficient estimates remain within their nominal physical ranges. Open-loop and closed-loop performance assessments, using a physics-based simulator and full-scale autonomous Indy racecar data, highlight Deep Dynamics as a promising approach for modeling racecar vehicle dynamics.
title Deep Dynamics: Vehicle Dynamics Modeling with a Physics-Constrained Neural Network for Autonomous Racing
topic Robotics
Artificial Intelligence
Machine Learning
I.2.9; I.6
url https://arxiv.org/abs/2312.04374