Motion planning for highly automated road vehicles with a hybrid approach using nonlinear optimization and artificial neural networks

Ferenc Hegedüs, Tamás Bécsi, Szilárd Aradi, P. Gáspár

Research output: Contribution to journalArticle

Abstract

Over the last decade, many different algorithms were developed for the motion planning of road vehicles due to the increasing interest in the automation of road transportation. To be able to ensure dynamical feasibility of the planned trajectories, nonholonomic dynamics of wheeled vehicles must be considered. Nonlinear optimization based trajectory planners are proven to satisfy this need, however this happens at the expense of increasing computational effort, which jeopardizes the real-time applicability of these methods. This paper presents an algorithm which offers a solution to this problematic with a hybrid approach using artificial neural networks (ANNs). First, a nonlinear optimization based trajectory planner is presented which ensures the dynamical feasibility with the model-based prediction of the vehicle’s motion. Next, an artificial neural network is trained to reproduce the behavior of the optimization based planning algorithm with the method of supervised learning. The generation of training data happens off-line, which eliminates the concerns about the computational requirements of the optimization-based method. The trained neural network then replaces the original motion planner in on-line planning tasks which significantly reduces computational effort and thus run-time. Furthermore, the output of the network is supervised by the model based motion prediction layer of the original optimization-based algorithm and can thus always be trusted. Finally, the performance of the hybrid method is benchmarked with computer simulations in terms of dynamical feasibility and run-time and the results are investigated. Examinations show that the computation time can be significantly reduced while maintaining the feasibility of resulting vehicle motions.

Original languageEnglish
Pages (from-to)148-160
Number of pages13
JournalStrojniski Vestnik/Journal of Mechanical Engineering
Volume65
Issue number3
DOIs
Publication statusPublished - Jan 1 2019

Fingerprint

Motion planning
Neural networks
Trajectories
Planning
Supervised learning
Automation
Computer simulation

Keywords

  • Artificial neural networks
  • Automated driving
  • Motion planning
  • Nonlinear optimization
  • Trajectory planning
  • Vehicle control

ASJC Scopus subject areas

  • Mechanics of Materials
  • Mechanical Engineering

Cite this

Motion planning for highly automated road vehicles with a hybrid approach using nonlinear optimization and artificial neural networks. / Hegedüs, Ferenc; Bécsi, Tamás; Aradi, Szilárd; Gáspár, P.

In: Strojniski Vestnik/Journal of Mechanical Engineering, Vol. 65, No. 3, 01.01.2019, p. 148-160.

Research output: Contribution to journalArticle

@article{ccd63d5ffd26455ea634499df8cf6dde,
title = "Motion planning for highly automated road vehicles with a hybrid approach using nonlinear optimization and artificial neural networks",
abstract = "Over the last decade, many different algorithms were developed for the motion planning of road vehicles due to the increasing interest in the automation of road transportation. To be able to ensure dynamical feasibility of the planned trajectories, nonholonomic dynamics of wheeled vehicles must be considered. Nonlinear optimization based trajectory planners are proven to satisfy this need, however this happens at the expense of increasing computational effort, which jeopardizes the real-time applicability of these methods. This paper presents an algorithm which offers a solution to this problematic with a hybrid approach using artificial neural networks (ANNs). First, a nonlinear optimization based trajectory planner is presented which ensures the dynamical feasibility with the model-based prediction of the vehicle’s motion. Next, an artificial neural network is trained to reproduce the behavior of the optimization based planning algorithm with the method of supervised learning. The generation of training data happens off-line, which eliminates the concerns about the computational requirements of the optimization-based method. The trained neural network then replaces the original motion planner in on-line planning tasks which significantly reduces computational effort and thus run-time. Furthermore, the output of the network is supervised by the model based motion prediction layer of the original optimization-based algorithm and can thus always be trusted. Finally, the performance of the hybrid method is benchmarked with computer simulations in terms of dynamical feasibility and run-time and the results are investigated. Examinations show that the computation time can be significantly reduced while maintaining the feasibility of resulting vehicle motions.",
keywords = "Artificial neural networks, Automated driving, Motion planning, Nonlinear optimization, Trajectory planning, Vehicle control",
author = "Ferenc Heged{\"u}s and Tam{\'a}s B{\'e}csi and Szil{\'a}rd Aradi and P. G{\'a}sp{\'a}r",
year = "2019",
month = "1",
day = "1",
doi = "10.5545/sv-jme.2018.5802",
language = "English",
volume = "65",
pages = "148--160",
journal = "Strojniski Vestnik/Journal of Mechanical Engineering",
issn = "0039-2480",
publisher = "Assoc. of Mechanical Eng. and Technicians of Slovenia",
number = "3",

}

TY - JOUR

T1 - Motion planning for highly automated road vehicles with a hybrid approach using nonlinear optimization and artificial neural networks

AU - Hegedüs, Ferenc

AU - Bécsi, Tamás

AU - Aradi, Szilárd

AU - Gáspár, P.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Over the last decade, many different algorithms were developed for the motion planning of road vehicles due to the increasing interest in the automation of road transportation. To be able to ensure dynamical feasibility of the planned trajectories, nonholonomic dynamics of wheeled vehicles must be considered. Nonlinear optimization based trajectory planners are proven to satisfy this need, however this happens at the expense of increasing computational effort, which jeopardizes the real-time applicability of these methods. This paper presents an algorithm which offers a solution to this problematic with a hybrid approach using artificial neural networks (ANNs). First, a nonlinear optimization based trajectory planner is presented which ensures the dynamical feasibility with the model-based prediction of the vehicle’s motion. Next, an artificial neural network is trained to reproduce the behavior of the optimization based planning algorithm with the method of supervised learning. The generation of training data happens off-line, which eliminates the concerns about the computational requirements of the optimization-based method. The trained neural network then replaces the original motion planner in on-line planning tasks which significantly reduces computational effort and thus run-time. Furthermore, the output of the network is supervised by the model based motion prediction layer of the original optimization-based algorithm and can thus always be trusted. Finally, the performance of the hybrid method is benchmarked with computer simulations in terms of dynamical feasibility and run-time and the results are investigated. Examinations show that the computation time can be significantly reduced while maintaining the feasibility of resulting vehicle motions.

AB - Over the last decade, many different algorithms were developed for the motion planning of road vehicles due to the increasing interest in the automation of road transportation. To be able to ensure dynamical feasibility of the planned trajectories, nonholonomic dynamics of wheeled vehicles must be considered. Nonlinear optimization based trajectory planners are proven to satisfy this need, however this happens at the expense of increasing computational effort, which jeopardizes the real-time applicability of these methods. This paper presents an algorithm which offers a solution to this problematic with a hybrid approach using artificial neural networks (ANNs). First, a nonlinear optimization based trajectory planner is presented which ensures the dynamical feasibility with the model-based prediction of the vehicle’s motion. Next, an artificial neural network is trained to reproduce the behavior of the optimization based planning algorithm with the method of supervised learning. The generation of training data happens off-line, which eliminates the concerns about the computational requirements of the optimization-based method. The trained neural network then replaces the original motion planner in on-line planning tasks which significantly reduces computational effort and thus run-time. Furthermore, the output of the network is supervised by the model based motion prediction layer of the original optimization-based algorithm and can thus always be trusted. Finally, the performance of the hybrid method is benchmarked with computer simulations in terms of dynamical feasibility and run-time and the results are investigated. Examinations show that the computation time can be significantly reduced while maintaining the feasibility of resulting vehicle motions.

KW - Artificial neural networks

KW - Automated driving

KW - Motion planning

KW - Nonlinear optimization

KW - Trajectory planning

KW - Vehicle control

UR - http://www.scopus.com/inward/record.url?scp=85063458778&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063458778&partnerID=8YFLogxK

U2 - 10.5545/sv-jme.2018.5802

DO - 10.5545/sv-jme.2018.5802

M3 - Article

AN - SCOPUS:85063458778

VL - 65

SP - 148

EP - 160

JO - Strojniski Vestnik/Journal of Mechanical Engineering

JF - Strojniski Vestnik/Journal of Mechanical Engineering

SN - 0039-2480

IS - 3

ER -