LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction

Autonomous driving can obtain accurate perception and reliable motion prediction with the support of multi-modal fusion. Recently, there has been growing interest in leveraging features from various onboard sensors to enhance the primary stages of autonomous driving. This paper proposes <italic&g...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Yasser H. Khalil, Hussein T. Mouftah
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Autonomous driving deep learning motion prediction multi-modal fusion perception residual image Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/1646adcd5a5f43debca63ab9287b24e2
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:1646adcd5a5f43debca63ab9287b24e2
record_format	dspace
spelling	oai:doaj.org-article:1646adcd5a5f43debca63ab9287b24e22021-11-09T00:00:50ZLiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction2169-353610.1109/ACCESS.2021.3123169https://doaj.org/article/1646adcd5a5f43debca63ab9287b24e22021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9585703/https://doaj.org/toc/2169-3536Autonomous driving can obtain accurate perception and reliable motion prediction with the support of multi-modal fusion. Recently, there has been growing interest in leveraging features from various onboard sensors to enhance the primary stages of autonomous driving. This paper proposes <italic>LiCaNext</italic> to capture additional accuracy advancements in joint perception and motion prediction while maintaining real-time requirements. <italic>LiCaNext</italic> is the <italic>next</italic> version of LiCaNet, which fuses LIDAR data in both bird’s-eye view (BEV) and range view (RV) representations with a camera image. In contrast to LiCaNet, we introduce sequential range residual images into the multi-modal fusion network to further improve performance, with minimal increase in inference time. Employing sequential range residual images has a substantial direct impact on motion prediction and positively influences perception. We provide an extensive evaluation on the public nuScenes dataset. Our experiments show that incorporating sequential range residuals secures significant performance gain, with monotonic progress for a larger number of exploited residuals.Yasser H. KhalilHussein T. MouftahIEEEarticleAutonomous drivingdeep learningmotion predictionmulti-modal fusionperceptionresidual imageElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 146244-146255 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Autonomous driving deep learning motion prediction multi-modal fusion perception residual image Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Autonomous driving deep learning motion prediction multi-modal fusion perception residual image Electrical engineering. Electronics. Nuclear engineering TK1-9971 Yasser H. Khalil Hussein T. Mouftah LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
description	Autonomous driving can obtain accurate perception and reliable motion prediction with the support of multi-modal fusion. Recently, there has been growing interest in leveraging features from various onboard sensors to enhance the primary stages of autonomous driving. This paper proposes <italic>LiCaNext</italic> to capture additional accuracy advancements in joint perception and motion prediction while maintaining real-time requirements. <italic>LiCaNext</italic> is the <italic>next</italic> version of LiCaNet, which fuses LIDAR data in both bird’s-eye view (BEV) and range view (RV) representations with a camera image. In contrast to LiCaNet, we introduce sequential range residual images into the multi-modal fusion network to further improve performance, with minimal increase in inference time. Employing sequential range residual images has a substantial direct impact on motion prediction and positively influences perception. We provide an extensive evaluation on the public nuScenes dataset. Our experiments show that incorporating sequential range residuals secures significant performance gain, with monotonic progress for a larger number of exploited residuals.
format	article
author	Yasser H. Khalil Hussein T. Mouftah
author_facet	Yasser H. Khalil Hussein T. Mouftah
author_sort	Yasser H. Khalil
title	LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
title_short	LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
title_full	LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
title_fullStr	LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
title_full_unstemmed	LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction
title_sort	licanext: incorporating sequential range residuals for additional advancement in joint perception and motion prediction
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/1646adcd5a5f43debca63ab9287b24e2
work_keys_str_mv	AT yasserhkhalil licanextincorporatingsequentialrangeresidualsforadditionaladvancementinjointperceptionandmotionprediction AT husseintmouftah licanextincorporatingsequentialrangeresidualsforadditionaladvancementinjointperceptionandmotionprediction
_version_	1718441363510394880

LiCaNext: Incorporating Sequential Range Residuals for Additional Advancement in Joint Perception and Motion Prediction

Ejemplares similares