Thomas Petzold is Head of Camera & Ultrasonic at Aumovio, where he is responsible for the global camera and ultrasonic business across the full value chain – from early innovation and business development through product strategy, development, industrialization, and series lifecycle management.
Petzold brings nearly two decades of leadership experience from Continental, where he held several senior roles within ADAS. As Head of Camera & Ultrasonic, he led global portfolios covering research, business case development, product management, system and software development, industrialization, and long-term series support. Previously, as Head of Radar Strategy and Base Programs, he was responsible for portfolio strategy, platform deployment, and base program management within the radar segment.
In earlier roles, Thomas headed Virtual Systems, where he established cutting-edge simulation and virtual validation environments for ADAS and automated driving, enabling automated function validation across global customer projects and significantly accelerating development cycles. His background also includes global responsibility for test and validation, as well as building and leading regional test organizations in Asia, giving him strong international and cross-cultural leadership experience.
He holds an MBA in international leadership and business management, a Diplom-Wirtschaftsingenieur (FH) in electrical engineering, and a Bachelor of engineering with distinction in electronic engineering.
DVN: Hi, Thomas. What does Aumovio’s camera portfolio look like?
Thomas Petzold: Our portfolio covers a broad range of cameras. A particular emphasis lies on camera heads (‘satellite cameras’) which simply record images and send them to a central processing unit. Their applications span from surround view and parking to driver assistance, and even full 360° vision as a camera belt around the vehicle for higher levels of autonomy. In addition to these, we offer satellite cameras with integrated processing circuitry support detection tasks for rear‑view and rear emergency brake assist. Through a joint venture with a Chinese SoC vendor, we also develop smart cameras for front‑looking applications. IR cameras in the short‑wave IR range complement our portfolio by enhancing visibility in challenging conditions such as nighttime or adverse weather. Finally, our offering includes cameras for in‑cabin sensing and driver monitoring.
DVN: What types of imagers and embedded image processing pipelines, hardware and software optimizations are to be found in Aumovio cameras?
Thomas Petzold: Our satellite cameras can host any imager, however we are usually using an imager from one of the four big imager vendors in the market. We focus on CMOS technology with a dynamic range of up to 140 dB, a pixel pitch ranging from 2.1 to 3.2 µm for satellite, and 4.2 µm in smart cameras with a pixel range from 1 to 8 megapixels. Imagers with higher resolution around 12 MP are in evaluation with customers. Most of our satellite cameras provide raw images. Smarter satellite cameras with processing circuitry inside offer image quality tuned images up to detection output for rear emergency braking. Our smart cameras offer full detection and fusion capabilities for tasks related to driver assistance systems.
We complement our cameras with a broad service portfolio, including simulation, integration, concepting, testing and rapid prototyping, and support our customers in image quality tuning and hardware-related software development.
DVN: What kinds of algorithms and system architectures do you use to cope with challenging conditions, such as low light and adverse weather?
Thomas Petzold: Our cameras are designed to meet the requirements and work seamless with all common CV-stack providers. The HDR of up to 140 dB provides reliable contrast for perception in low-light scenarios, that can be extended by IR illumination for more visibility in scenarios like FMVSS 127. An overall system is complemented by radars, where perception stacks leverage the fusion capabilities for reliable driving functions in difficult scenarios with adverse weather and light. In addition, we work on hardware-related software schemes that can enhance the contrast and reduces the noise for better detection and function performance.
DVN: What embedded AI stacks do you implement? What use cases are supported natively?
Thomas Petzold: Our cameras are designed to meet the requirements of common AI stacks and perceptions. The cameras come with tailored resolution, high quality and HDR, and can support multiple use cases like environment perception implemented in an AI stack natively. The performance of an AI stack in corner cases depends on a close alignment between the capabilities of both the cameras and the perception stack. In other words, our cameras and perception stacks form a great symbiose to meet function requirements in the automotive and non-automotive industry.
DVN: Are your cameras OTA-updateable? Which hardware parts are upgradeable?
Thomas Petzold: Our satellite cameras follow a highly scalable and modular hardware design. This allows us to provide hardware updates like changes in the lens stack or imager if wished by our customers. Updating camera parameters via OTA is possible if the corresponding central processing unit offers the option to write parameters into our satellite cameras.
DVN: What is your current standing by market segment?
Thomas Petzold: Our current focus is on all OEMs in the market for passenger and commercial vehicles worldwide. We are open to subtly explore the options for selling our cameras in non-automotive markets like robots, humanoid robots and drones.
DVN: Which market segments demonstrate the highest growth for your solutions, particularly regarding advanced technology?
Thomas Petzold: We see an increased need for satellite cameras in the automotive market for passenger vehicle during the next years, where satellite cameras provide images to central processing units providing 360° vision, perception or other related tasks to driving.
DVN: How do your products compare with competing offerings? What strategy defines your price-performance positioning?
Thomas Petzold: Our satellite cameras, services, and technology around provide an optimized sweet spot between costs and performance required by an OEM. They are also well designed to meet the requirements of leading AI stacks, and strike excellent balances between performance, costs and power consumption. Our modular and scalable design offers best choices for quickly reacting to market needs and meeting price and performance constraints.
DVN: What significant technological advancements do you foresee over the next three to five years in embedded vision systems?
Thomas Petzold: System availability in difficult scenarios is key to increase the user acceptance of driving functions providing support from visualization only, assisted and automated driving. Current technology advancements for new imagers and software point towards increased system availability, and our strategy defines close alignments and collaborations with vendors providing such promising technologies.
DVN: What R&D areas do you prioritize for progress toward higher levels of autonomy and integration into zonal architectures?
Thomas Petzold: We develop our cameras with an understanding that the performance of systems for higher-level autonomy begins when light enters the lens of a camera. In addition, we consider our cameras including the services provided as a crucial contributor to higher level of autonomy from a system perspective. This means that we need to put an equal emphasis on all R&D areas contributing to the physical hardware of the cameras, the software and services around to ensure that our todays and tomorrow’s cameras work in the entire system.
DVN: Thank you, Thomas, for providing detailed information about Aumovio’s cameras, technologies, and markets.
