Driving scenarios detection is an important aspect of the development of automated driving functions (ADF). Given the lack of publicly available datasets with driving scenario labels, we designed a toolchain for generating synthetic video datasets of driving scenarios, based on the OpenSCENARIO format, a well-established, public and vendor-independent standard. The experience reported in this paper shows the feasibility of a full end-to-end implementation of a workflow allowing designers to quickly create datasets for pre-training machine learning models. Video clips are recorded through a driving simulator which runs different sessions implementing variations of a pre-defined set of driving scenarios. The user specifies through a configuration file each parameter value range (e.g., vehicle speed, distance, weather conditions) that represent the intended variability within each scenario. Preliminary results show effectiveness of the approach and indicate directions on how to improve the system and reduce the need for human intervention in post-production.

Developing a Synthetic Dataset for Driving Scenarios

Bellotti F.;Berta R.;Capello A.;Cossu M.;De Gloria A.;Lazzaroni L.;Bonora S.
2022-01-01

Abstract

Driving scenarios detection is an important aspect of the development of automated driving functions (ADF). Given the lack of publicly available datasets with driving scenario labels, we designed a toolchain for generating synthetic video datasets of driving scenarios, based on the OpenSCENARIO format, a well-established, public and vendor-independent standard. The experience reported in this paper shows the feasibility of a full end-to-end implementation of a workflow allowing designers to quickly create datasets for pre-training machine learning models. Video clips are recorded through a driving simulator which runs different sessions implementing variations of a pre-defined set of driving scenarios. The user specifies through a configuration file each parameter value range (e.g., vehicle speed, distance, weather conditions) that represent the intended variability within each scenario. Preliminary results show effectiveness of the approach and indicate directions on how to improve the system and reduce the need for human intervention in post-production.
978-3-030-95497-0
978-3-030-95498-7
File in questo prodotto:
File Dimensione Formato  
Applepies Dataset 3.pdf

accesso chiuso

Descrizione: Contributo in atti di convegno
Tipologia: Documento in Post-print
Dimensione 373.94 kB
Formato Adobe PDF
373.94 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/1094515
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact