Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DKMGP: A Gaussian Process Approach to Multi-Task and Multi-Step Vehicle Dynamics Modeling in Autonomous Racing

Jingyun Ning, Madhur Behl

TL;DR

The results show that DKMGP is a scalable and efficient solution for vehicle dynamics modeling making it suitable for high-speed autonomous racing control.

Abstract

Autonomous racing is gaining attention for its potential to advance autonomous vehicle technologies. Accurate race car dynamics modeling is essential for capturing and predicting future states like position, orientation, and velocity. However, accurately modeling complex subsystems such as tires and suspension poses significant challenges. In this paper, we introduce the Deep Kernel-based Multi-task Gaussian Process (DKMGP), which leverages the structure of a variational multi-task and multi-step Gaussian process model enhanced with deep kernel learning for vehicle dynamics modeling. Unlike existing single-step methods, DKMGP performs multi-step corrections with an adaptive correction horizon (ACH) algorithm that dynamically adjusts to varying driving conditions. To validate and evaluate the proposed DKMGP method, we compare the model performance with DKL-SKIP and a well-tuned single-track model, using high-speed dynamics data (exceeding 230kmph) collected from a full-scale Indy race car during the Indy Autonomous Challenge held at the Las Vegas Motor Speedway at CES 2024. The results demonstrate that DKMGP achieves upto 99% prediction accuracy compared to one-step DKL-SKIP, while improving real-time computational efficiency by 1752x. Our results show that DKMGP is a scalable and efficient solution for vehicle dynamics modeling making it suitable for high-speed autonomous racing control.

DKMGP: A Gaussian Process Approach to Multi-Task and Multi-Step Vehicle Dynamics Modeling in Autonomous Racing

TL;DR

The results show that DKMGP is a scalable and efficient solution for vehicle dynamics modeling making it suitable for high-speed autonomous racing control.

Abstract

Autonomous racing is gaining attention for its potential to advance autonomous vehicle technologies. Accurate race car dynamics modeling is essential for capturing and predicting future states like position, orientation, and velocity. However, accurately modeling complex subsystems such as tires and suspension poses significant challenges. In this paper, we introduce the Deep Kernel-based Multi-task Gaussian Process (DKMGP), which leverages the structure of a variational multi-task and multi-step Gaussian process model enhanced with deep kernel learning for vehicle dynamics modeling. Unlike existing single-step methods, DKMGP performs multi-step corrections with an adaptive correction horizon (ACH) algorithm that dynamically adjusts to varying driving conditions. To validate and evaluate the proposed DKMGP method, we compare the model performance with DKL-SKIP and a well-tuned single-track model, using high-speed dynamics data (exceeding 230kmph) collected from a full-scale Indy race car during the Indy Autonomous Challenge held at the Las Vegas Motor Speedway at CES 2024. The results demonstrate that DKMGP achieves upto 99% prediction accuracy compared to one-step DKL-SKIP, while improving real-time computational efficiency by 1752x. Our results show that DKMGP is a scalable and efficient solution for vehicle dynamics modeling making it suitable for high-speed autonomous racing control.

Paper Structure

This paper contains 16 sections, 6 equations, 6 figures, 4 tables, 1 algorithm.

Table of Contents

  1. INTRODUCTION
  2. PROBLEM FORMULATION
  3. Single-track Model
  4. Extended Kinematic Model
  5. Problem Formulation
  6. METHODOLOGY
  7. Deep Kernel Learning
  8. Multi-Task Gaussian Process
  9. DKL Integration
  10. MULTI-STEP DKMGP
  11. EXPERIMENTS AND RESULTS
  12. Real-world Data Collection
  13. DKMGP Hyperparameters Tuning
  14. DKMGP Model Validation
  15. Comparison of DKMGP and Nonlinear Singe-track
  16. ...and 1 more sections

Figures (6)

  • Figure 1: The single-track vehicle model of the racecar, with all three base states $(v_x. v_y, \omega)$ highlighted.
  • Figure 2: The architecture of the DKMGP model.
  • Figure 3: Workflow of the multi-step approach for the DKMGP predictions.
  • Figure 4: (A) Data from overtaking competition, LVMS CES 2024. (B) LVMS track layout. (C) LiDAR point cloud data. (D) AV-21 sensor setup.
  • Figure 5: Comparison of DKMGP models across different fixed correction horizons.
  • ...and 1 more figures