AutoSens USA 2025 is set to deliver another groundbreaking agenda, bringing together top experts in ADAS, automotive perception, and sensing technology.
From deep dives into next-gen LiDAR and thermal imaging to discussions on regulatory frameworks and real-world deployment challenges, AutoSens 2025 offers a unique opportunity to stay ahead of the curve. Here’s a look at the must-attend sessions that will define the future of automotive sensing.
The Next Generation of Vehicle Diagnostics: Predictive Maintenance for Autonomous Vehicles
Sneha Shetiya, Staff Software Engineer, TORC Robotics
This presentation will explore the shortcomings of current On-Board Diagnostics (OBD) systems in meeting the evolving needs of autonomous vehicles. It will advocate for a proactive vehicle maintenance strategy as crucial for ensuring the safety and reliability of these complex systems. The existing OBD model primarily reports existing symptoms or faults, rather than predicting potential component failures. This highlights the importance of predictive maintenance, yet significant challenges remain in acquiring high-quality data through On-Board Monitoring (OBM). These challenges encompass developing standards for anomaly definition, creating effective algorithms, ensuring data privacy, and mitigating the risks of tampering and cybersecurity. Potential solutions to be discussed include advanced sensing technologies, Artificial Intelligence (AI), and Machine Learning (ML) for predictive analytics and attack detection. Furthermore, the presentation will emphasize the necessity of collaboration among vehicle manufacturers, suppliers, and regulators to develop and implement effective predictive maintenance strategies for sustainable solutions.
The Transition of ADAS to SDV
Douglas C. MacKenzie – Head of Intelligent Mobility Group, Mobis North America
Mobis has a full portfolio of sensors and perception systems for automotive applications covering the spectrum from manual driving through autonomous vehicles. Historically, these have been supplied as individual components with the software embedded in purpose-built ECUs. This presentation will cover Mobis’s on-going transition to zonal control units (ZCUs) replacing ECUs for software defined vehicles. The zonal control unit approach is a foundation for the vision of SDV where the OE can continuously provide new experiences and increased performance to users through software over-the-air updates, including the sensing and perception software running in the ZCU.
Applications for Automotive Remote Direct Memory Access (ARDMA) in a Software Defined Vehicle
Dr. Markus Jochim (left), Research Group Manager, Electrical and Software Architectur & Dr. Yuchen Zhou (right), Senior Researcher, Electrical and Software Architecture, GM R&D
Software Defined Vehicle Architectures centralize the implementations of low latency control algorithms as well ADAS features in a high-performance Central Compute Unit. Centralization leads to longer data paths and increases end-to-end latency on paths between sensors, the Central Compute Unit, and actuators with multiple time-consuming handovers between software tasks and hardware components. We introduce ARDMA developed by General Motors Research and Development and inspired by techniques used in Data Centers as a way for minimizing the end-to-end latency of control traffic and we show how ADAS sensor applications can benefit from this technology. Our ARDMA implementation in hardware on an FPGA demonstrates significant reductions in latency and computational overhead when compared to more traditional implementations and architectures.
Balancing Performance and Efficiency: Right-Sizing ADAS Sensor Suites Through Sensor Fusion
Elizabeth Kao, Director of Product Management for ADAS, Bosch
As the automotive industry pushes forward with ADAS and autonomous driving, achieving the delicate balance between safety and system efficiency is more crucial than ever. While the performance of sensors continues to improve, the increasing volume of data these sensors generate brings with it greater complexity, higher computational demands, and escalating costs. But do we really need to equip every vehicle with the most expensive LiDAR, radar, and cameras to enable autonomous features? Or is there a smarter, more efficient approach to designing sensor suites?
Join Elizabeth Kao, Director of Product Management for ADAS at Bosch for her keynote at AutoSens USA as she explores how sensor fusion can “right-size” ADAS sensor suites to optimize performance without compromising safety. By combining camera and high-precision radar sensors and leveraging machine learning, we can significantly enhance safety systems, such as Automatic Emergency Braking (AEB), even in challenging conditions like the dark of the night. Discover how this innovative approach can extend detection range and improve system efficiency, paving the way for smarter, more cost-effective ADAS solutions.
Designing a Virtual Verification Machine for Safety Assurance of Continuously Evolving Automated Driving SystemsAdd Your Heading Text Here
Håkan Sivencrona, Senior Technical Leader – AD/ADAS Safety, Volvo
Continuous integration and deployment of software including machine learning components. DevOps is a standard practice in IT and cloud systems to reduce the time between Development and Operations, enabling frequent releases. The practice is closely related to agile development with abilities to quickly respond to changes in requirements.
Safety engineering practices placing stringent emphasis on design for safety, and Verification and Validation (V&V) – characterized by very long and slow cycles in software and data updates. A key reason for this is that safety is a property that depends on the behavior of an integrated system as-a-whole in its environment, and hence it is not enough to verify only the software, requiring costly and complicated V&V efforts. Academically, ”SafeOps” or “SafeDevOps” represent recent research themes in safety conferences and exemplified with recent international projects such as the large AutoDevSafeOps German project, with emphasis on systems engineering methods including V&V methodologies.
AI and Connected Intelligence: Harnessing the Full Potential of Software-Defined System
Javed Khan, President of Software and Advanced Safety and User Experience, Aptiv
The future of mobility lies in software-defined fleets that are intelligent, connected, and user-centric. This fireside chat with Javed Khan, President of Software and Advanced Safety and User Experience, Aptiv, we discuss how AI and connectivity are unlocking the full potential of vehicles and connected devices. We will explore how Aptiv is leveraging AI/ML to enhance vehicle intelligence, from advanced driver-assistance systems (ADAS) to in-cabin experiences. The conversation will also delve into examine the critical role of secure and reliable connectivity, including 5G, in enabling seamless data flow and over-the-air updates that continuously improve vehicle performance and user experience.
Interested in in-cabin monitoring technology?
With a pass to AutoSens USA, you’ll also get full access to our co-located sister event, InCabin. See the key topics for InCabin here >>
