Poziv na predavanje "Perception Error Modeling for Autonomous Driving"

Abstract

This presentation begins with an overview of perception systems in autonomous vehicles, highlighting their main components, sensing modalities, and integration challenges. It then focuses on virtual testing for autonomous driving, which is essential for safety assessment but still limited by how sensing and perception are represented in simulation.

To address this gap, we introduce Perception Error Models (PEMs), a simulation component that enables analysis of perception errors without explicitly modeling the sensors. A data-driven procedure for parametric modeling is proposed and evaluated using Apollo, an open-source driving stack, and the nuScenes dataset. The PEMs are further implemented in SVL, an open-source vehicle simulator, and tested with camera, LiDAR, and camera-LiDAR configurations. The results highlight limitations in current evaluation metrics and demonstrate how PEM-based virtual tests can improve the understanding of perception-related safety risks in autonomous vehicles.

Biography

Dr. Justin Dauwels is an Associate Professor at the TU Delft (Signals and Systems, Department of Microelectronics), and serves as co-Director of the Safety and Security Institute at the TU Delft. He was an Associate Professor of the School of Electrical and Electronic Engineering at the Nanyang Technological University (NTU) in Singapore till the end of 2020. At the TU Delft, he serves as scientific lead of the Model-Driven Decisions Lab (MoDDL), a first lab for the Knowledge Building program between the police and the TU Delft. He also serves as Chairperson of the EE Board of Studies at the TU Delft, and is a board member of the Netherlands Institute for Research on ICT.

His research interests are in data analytics with applications to intelligent transportation systems, autonomous systems, and analysis of human behaviour and physiology. He obtained his PhD degree in electrical engineering at the Swiss Polytechnical Institute of Technology (ETH) in Zurich in December 2005. Moreover, he was a postdoctoral fellow at the RIKEN Brain Science Institute (2006-2007) and a research scientist at the Massachusetts Institute of Technology (2008-2010).

He has been elected as IEEE SPS 2024 Distinguished Lecturer. He served as Chairman of the IEEE CIS Chapter in Singapore from 2018 to 2020, and served as Associate Editor of the IEEE Transactions on Signal Processing (2018 - 2023), and serves currently as Associate Editor (2021-2023) and Subject Editor (since 2023) of the Elsevier journal Signal Processing, Area Editor C&F for the IEEE Signal Processing Magazine (since 2023), member of the Editorial Advisory Board of the International Journal of Neural Systems (since 2021), and organizer of IEEE conferences and special sessions. He was also Elected Member of the IEEE Signal Processing Theory and Methods Technical Committee and IEEE Biomedical Signal Processing Technical Committee (both in 2018-2023), and is currently Elected Member of the IEEE Machine Learning for Signal Processing Technical Committee and the IEEE Emerging Transportation Technology Testing (ET3) Technical Committee. He has been a JSPS postdoctoral fellow (2007), a BAEF fellow (2008), a Henri-Benedictus Fellow of the King Baudouin Foundation (2008), and a JSPS invited fellow (2010, 2011). His research team has won several best paper awards at international conferences and journals.

His research on intelligent transportation systems has been featured by the BBC, Straits Times, Lianhe Zaobao, Channel 5, and numerous technology websites. Besides his academic efforts, the team of Dr. Justin Dauwels also collaborates intensely with local start-ups, SMEs, and agencies, in addition to MNCs, in the field of data-driven transportation, logistics, and medical data analytics. His academic lab has spawned four startups across a range of industries, ranging from AI for healthcare to autonomous vehicles.