Driving a scalable, consumer-centric vision in the mobility industry, vehicles of thefuture will always be connected and differentiated by software. Advancements in software, hardware and their interaction are expanding the boundaries of performance, providing the foundation for next-generation cars. But the same technology that will make this vision a reality also presents new challenges.
Over time, automotive electronic systems have become the most complex element of vehicle architecture. Software in today’s cars can contain more than
100 million lines of code. In comparison, software for a modern commercial aircraft contains “only” ~10 million lines of code. Advanced System-on-Chips (SoCs) are populated with billions of transistors, a figure that has doubled annually over the past three decades. New architectures are being designed to allow for higher performance without increasing power or area. Autonomous driving and electrification are accelerating this trend. To reduce complexity and interdependencies, Automakers are moving toward consolidating electronics into domain controllers.
Reliable implementation of these advanced technologies must meet stringent zero-downtime requirements, while accommodating unpredictable environmental and operational conditions. Fleet managers need to guarantee the availability and seamless experience of these complex, always-on and always-connected systems. Yet, current monitoring and diagnostic practices lack the data resolution and context to meet these requirements.
This paper outlines an innovative approach to Predictive and Preventive Maintenance (PPM) by extracting and analyzing deep data from within SoC devices using machine learning algorithms that allow manufacturers to:
Leveraging this data allows for enhanced PPM practices to meet the surging safety and reliability requirements of today’s software-enabled vehicles, while reducing maintenance costs. The approach utilizes Over-the-Air (OTA) technology and advanced device health monitoring capabilities to collect vehicle data and apply software updates to a pre-defined subset of the fleet.