The Role of Cameras and Photonics for Smart Cars
Full day workshop on Sep 8th, 2025 @ ESSERC 2025 TU Munich
Organizers
Cedric Tubert (STMicroelectronics, FR)
Daniele Perenzoni (Sony, IT)
This workshop explores cutting-edge developments in automotive vision systems, highlighting the integration of advanced cameras and photonic technologies. We present significant advancements in Automotive High Dynamic Range Imaging specifically designed for High Temperature conditions, addressing one of the industry's most challenging operational environments. The session showcases innovative wafer-scale micro-optics and meta-surfaces that are revolutionizing both imaging and illumination applications. Attendees will gain insights into the evolution of next generation CMOS image sensors for smart cars and for Driver and Occupancy Monitoring systems. The workshop also examines hardware accelerators enabling low-latency event-based vision processing, critical for real-time decision. Finally, we address the integration challenges in 'Photonics on the Road,' exploring practical hurdles and solutions for implementing these technologies in self-driving vehicles. These innovations collectively demonstrate the essential role of photonics and imaging systems in creating safer, more efficient autonomous transportation.
Program
09:30 - 10:15
Automotive High Dynamic Range Imaging in High Temperature Conditions
Tomas Geurts (Omnivision, BE)
The talk will cover High Dynamic Range (HDR) requirements in ADAS and In-Cabin automotive imaging applications. The importance and relevance of performance at high temperature will be explained. The talk will highlight fundamental limitations of low-light and HDR performance at elevated temperatures which is an important aspect in automotive imaging but often under-illuminated in publications.
10:15 - 11:00
Past and Future of CMOS Image Sensors in Automotive Industry
Yorito Sakano (Sony Semiconductor Solutions, JP)
Business motivation is essential for the evolution of semiconductor devices. The larger the market, the faster the technology evolves. The first iPhone was born in 2007, and the back-illuminated image sensor, an epoch-making event for CMOS image sensors, was introduced in 2009. With technical breakthroughs and business motivations coming together almost simultaneously, CMOS image sensors have undergone a dramatic technological evolution over the past decade or so. Similarly, automotive CMOS image sensors have recently undergone a unique evolution in the competitive axis of high dynamic range (HDR), supported by business motivation such as the evolution of Advanced Driver-Assistance Systems (ADAS) and the efforts toward the practical application of Autonomous Driving (AD). Let me overview the recent evolution of automotive CMOS image sensors and discuss the direction of future evolution.
11:00 - 11:30
Coffee break
11:30 - 12:15
Wafer Scale Micro-optics and Meta-surfaces for Applications in Imaging and Illumination
Falk Eilenberger (Fraunhofer, DE)
Micro- and nanooptical systems game-changers in our ability to manipulate light. Nanooptical systems, frequently called meta-surfaces, allow to access all degrees of freedom of the optical fields, such as spectral properties, its polarization, and its phase next to its intensity, which is classically addressed in imaging systems. Nano- and microoptical systems allow to introduce massive parallelization in optical systems, breaking virtually any commonly known design rules both for imaging as well as for illumination systems. Harnessing these degrees of freedom is, however, a grand challenge in terms of design, engineering, and cost scaling. In the talk I shall highlight how wafer scale fabrication techniques can be utilized to overcome these issues, if the entire process chain from design to the final application can be tailored to the specific requirements of the optical task at hand. I shall do so by highlighting a variety of applications and projects, in which wafer scale nanooptics have played a crucial role, from optics for satellite missions all the way to illumination systems for mobility solutions.
12:15 - 13:00
CMOS Image Sensors for Driver and Occupancy Monitoring Solutions
Jerome Chossat and Pierre Malinge (STMicroelectronics, FR)
Automotive applications require high-performance and cost-effective sensors. Considering these constraints, we present a novel pixel architecture capable of both rolling and global shutter imaging. Utilizing a non-Bayer CFA pattern, it captures both RGB and near-infrared images. A specific ASIL pixel design ensures a comprehensive integrity check of the sensor. The latter is connected to a logic circuit through a 3D Cu-to-Cu hybrid bonding process, providing state-of-the-art on-chip data processing and interfacing. Such a sensor is ideally suited for driver monitoring systems while enabling the integration of advanced multimedia features. Indeed, on top of the pixel and readout quality requirements, CMOS mage sensors for Driver and Occupancy Monitoring solutions are bringing a lot of challenges on the digital side too. They may contain quite complex signal processing for properly dealing with various non-Bayer CFA and manage IR content, they must integrate automotive safety capabilities, must be efficiently protected against malicious attackers aimed at tampering their functionalities, and must prevent usage of counterfeit components. In addition, all this must be done under aggressive cost constraint and stringent power constraints but also be developed in conformance with road vehicles functional safety (ISO26262), and Road vehicles Cybersecurity engineering (ISO21434).
13:00 - 14:00
Lunch
14:00 - 14:45
Hardware Accelerators for Low-latency Event-based Vision
Charlotte Frenkel (TU Delft, NL)
From optical flow to high-speed particle counting, event-based cameras emerge as an enabler for low-latency vision applications. They capture temporal contrast changes as a stream of events, which are generated on a per-pixel basis and at a temporal resolution of a few microseconds. However, there is currently a lack of hardware support for event-based processing workloads that generate updated predictions within microseconds. This talk will cover emerging developments in this area, from dynamic graph neural networks to digital in-memory computing for spiking neural networks.
14:45 - 15:30
Photonics on the Road: Navigating the Integration Hurdles in Self-Driving Cars
Christoph Parl (Valeo, DE)
Valeo is at the forefront of the autonomous driving revolution, providing a comprehensive suite of sensors - cameras, RADARs, ultrasonics, microphones, and LiDARs - that enable self-driving capabilities. This keynote will explore how Valeo's technology is driving the transition from manual to fully autonomous vehicles. A key focus will be on vehicle integration: the art of seamlessly embedding these sensors into the vehicle's design. This requires balancing function-driven design, ensuring optimal sensor performance, with emotion-driven design, creating desirable and engaging vehicles. The presentation will highlight the diverse sensors required for autonomy, with a focus on LiDARs due to complexity. Crucially, we will examine the challenges and solutions surrounding sensor mounting positions. Optimal placement is vital, considering each sensor's needs, environmental factors, and cleaning requirements. Finally, we'll explore how solid-state technology can help vehicle integration to enable more compact and robust solutions for a large-scale rollout of self-driving functions.
15:30 - 16:00
Coffee break
16:00 - 16:45
Final discussion and closing
No comments:
Post a Comment
All comments are moderated to avoid spam and personal attacks.