Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models
Prediction of surrounding vehicles accurately is an essential prerequisite for safe autonomous driving. Trajectory prediction methods can be classified into physics-, maneuver-, or learning-based methods. Learning-based methods have been studied extensively in recent years because it effectively exp...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2021-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9441017/ |
_version_ | 1811210779423145984 |
---|---|
author | Gihoon Kim Dongchan Kim Yoonyong Ahn Kunsoo Huh |
author_facet | Gihoon Kim Dongchan Kim Yoonyong Ahn Kunsoo Huh |
author_sort | Gihoon Kim |
collection | DOAJ |
description | Prediction of surrounding vehicles accurately is an essential prerequisite for safe autonomous driving. Trajectory prediction methods can be classified into physics-, maneuver-, or learning-based methods. Learning-based methods have been studied extensively in recent years because it effectively exploits the road information and interactions among vehicles. However, learning-based methods perform poorly in unseen environments that were not considered during training and provide unreasonable results such as inconsistent trajectories according to road geometry. In this paper, to address this problem, a hybrid model that combines a learning-based model with physics- and maneuver-based models according to their uncertainties is proposed. The deep ensemble technique is also used to estimate the uncertainty of the learning-based method. Because the deep ensemble tends to show a large variance in unseen environments, this method is used to determine whether to use a hybrid model. The proposed method is trained and validated using the Lyft l5 dataset, the real environment vehicle driving data containing several types of intersections. |
first_indexed | 2024-04-12T05:00:11Z |
format | Article |
id | doaj.art-6ff927cad09b44519964e7ed85d675cd |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-04-12T05:00:11Z |
publishDate | 2021-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-6ff927cad09b44519964e7ed85d675cd2022-12-22T03:47:00ZengIEEEIEEE Access2169-35362021-01-019787157872310.1109/ACCESS.2021.30839189441017Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple ModelsGihoon Kim0https://orcid.org/0000-0002-1631-7024Dongchan Kim1https://orcid.org/0000-0002-8251-7104Yoonyong Ahn2https://orcid.org/0000-0001-8934-0121Kunsoo Huh3https://orcid.org/0000-0002-7179-7841Department of Automotive Engineering, Hanyang University, Seoul, South KoreaDepartment of Automotive Engineering, Hanyang University, Seoul, South KoreaDepartment of Automotive Engineering (Automotive-Computer Convergence), Hanyang University, Seoul, South KoreaDepartment of Automotive Engineering, Hanyang University, Seoul, South KoreaPrediction of surrounding vehicles accurately is an essential prerequisite for safe autonomous driving. Trajectory prediction methods can be classified into physics-, maneuver-, or learning-based methods. Learning-based methods have been studied extensively in recent years because it effectively exploits the road information and interactions among vehicles. However, learning-based methods perform poorly in unseen environments that were not considered during training and provide unreasonable results such as inconsistent trajectories according to road geometry. In this paper, to address this problem, a hybrid model that combines a learning-based model with physics- and maneuver-based models according to their uncertainties is proposed. The deep ensemble technique is also used to estimate the uncertainty of the learning-based method. Because the deep ensemble tends to show a large variance in unseen environments, this method is used to determine whether to use a hybrid model. The proposed method is trained and validated using the Lyft l5 dataset, the real environment vehicle driving data containing several types of intersections.https://ieeexplore.ieee.org/document/9441017/Trajectory predictionphysics-based modelmaneuver-based modeldeep ensembleuncertaintyweighted integrated model |
spellingShingle | Gihoon Kim Dongchan Kim Yoonyong Ahn Kunsoo Huh Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models IEEE Access Trajectory prediction physics-based model maneuver-based model deep ensemble uncertainty weighted integrated model |
title | Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models |
title_full | Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models |
title_fullStr | Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models |
title_full_unstemmed | Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models |
title_short | Hybrid Approach for Vehicle Trajectory Prediction Using Weighted Integration of Multiple Models |
title_sort | hybrid approach for vehicle trajectory prediction using weighted integration of multiple models |
topic | Trajectory prediction physics-based model maneuver-based model deep ensemble uncertainty weighted integrated model |
url | https://ieeexplore.ieee.org/document/9441017/ |
work_keys_str_mv | AT gihoonkim hybridapproachforvehicletrajectorypredictionusingweightedintegrationofmultiplemodels AT dongchankim hybridapproachforvehicletrajectorypredictionusingweightedintegrationofmultiplemodels AT yoonyongahn hybridapproachforvehicletrajectorypredictionusingweightedintegrationofmultiplemodels AT kunsoohuh hybridapproachforvehicletrajectorypredictionusingweightedintegrationofmultiplemodels |