Combining Silicon Data Capture and Edge Processing Offers a Promising Solution for Automotive OEMs

Jasmin Mulaosmanovic, product management director at BlackBerry, co-authored this article.

In the world of electric (EV) and software-defined vehicles (SDV), a key challenge and opportunity besets original equipment manufacturers (OEM): that of effectively harnessing in-vehicle and silicon data to improve user experience and quite literally drive business benefits. While collecting telemetry data is straightforward enough, the difficulty lies in achieving the contextual understanding to transform it into valuable insights.

For example, OEMs generally understand the negative impact of extreme temperatures on vehicle electronics, but not at a holistic level. In addition to temperature, silicon variables such as timing margins, voltage, electromigration due to activity along with physical and environmental factors such as driving speed, driver assist features, external temperature, state of charge, usage of in-vehicle infotainment all play a role in the life expectancy of an automotive chip.

Consider the Arizona heatwave of July 2023, where the average temperature approached 103°F. Synopsys modeling indicated that within a 45-day window under such conditions, silicon could lose up to 10% of its useful life.  This carries significant implications given the potential for multiple bouts of extreme heat in certain regions. Monitoring silicon temperature in the context of a given environment is a way to track the stress that can lead to unexpected system failures. In the above case, while ambient temperature can lead to damage, silicon junction temperature can be 2x based on the chip activity or the structural cooling design of the system. Thus, observation and early diagnosis is crucial to avert the breakdown of electronic automotive components such as advanced driver assistance systems (ADAS) and uphold safety as well as wider functionality.

Taking this scenario a step further, let's imagine an EV owner is charging their car at a station in sweltering heat. They are watching a documentary  while their child is in the back seat playing a video game on the rear seat infotainment unit while the air conditioning is blasting. Suddenly, a message flashes: “Your vehicle infotainment unit is overheating. Please reduce usage to cool it down.”  After this warning message, additional details appear on the screen for the vehicle owner to educate themselves that reducing gaming or video playback will help cool down the infotainment unit and extend the life of their vehicle semiconductor components along with how to prevent future occurrences. The message is based on data collected from the vehicle and analyzed in real time.

Using anonymized diagnostic telemetry data from tens of thousands of cars, it becomes possible to offer proactive, specific guidance to owners based on their current usage of the vehicle, suggest preventative maintenance that might need to be done to their specific car, and provide regional warnings based on weather forecasts to help prevent silicon damage and extend their vehicle's lifetime. 

As of today, there is no single company that can provide the support, insights, and technology necessary for such a solution, but BlackBerry QNX and Synopsys are looking to team up to change that. BlackBerry IVY® is an in-vehicle data platform that incorporates edge processing and plans to enable a more proactive future for software-defined vehicles—one that can incorporate the advanced silicon health monitoring prognostics provided by Synopsys. 

Continuous Improvement through Edge Analytics

Let’s first examine the BlackBerry IVY platform, which enables automakers and their partners to innovate more effectively. Co-developed with Amazon Web Services (AWS), BlackBerry IVY software abstracts vehicle signals, enabling processing at the edge and cloud-controlled access to vehicle data. Car makers and software developers benefit from quicker development times, data connectivity savings, and optimized data processing using in-vehicle machine learning (ML).

The flow of the BlackBerry IVY Platform is as follows: 

• Data is collected from an SDV’s various sensors and systems including vehicle diagnostics, ADAS, body, cabin, and cameras. 
• This data is processed in real-time using edge computing capabilities within the vehicle. This includes data filtering, normalization, and aggregation. 
• Machine learning algorithms and analytics extract insights from the data related to vehicle performance and health, driver recognition and monitoring, and environmental conditions.  
• These IVY insights enable OEMs to rapidly experiment, innovate, and deploy new features over the lifecycle of the SDV in the areas of vehicle lifecycle and operations (e.g., predictive maintenance and insurance), e-commerce and monetization (e.g., in-vehicle payment), in-cabin experiences (e.g., personalization), and safety and security (e.g., crash analytics and driver authentication).

While BlackBerry IVY’s strengths lie in its processing of data within the vehicle, its connectivity to the cloud further boosts SDV value as it can share data and insights to the cloud as well as receive new logic and configurations. Additionally, it can input data from the cloud to enhance in-vehicle algorithms.

Software-Defined Vehicles Insights Through Combined Excellence

Figure 1: While hardware systems and architecture vary, real-time SLM silicon health data can be shared with an analytics engine like IVY for contextual insights for display within the vehicle or for cloud analytics.

A Future of Prolonged Vehicle Life

Ultimately, effective silicon health monitoring combined with other advanced vehicle diagnostics can prolong the useful lifetime of a vehicle. Owners now expect to run their vehicles for longer periods, with the average recently extended from 10 to 12 years. 

As SDVs move to more advanced architectures and high-performance compute (HPC) nodes with longer periods of operation, a holistic understanding of silicon performance in the automotive field is becoming vital. Effective monitoring today sets the foundation for greater efficiency in the future. Together with BlackBerry IVY, Synopsys can bring silicon-level data that will enable detailed behavior modeling over time, covering all aspects of SLM — a compelling proposition for a business sector defined by complexity and variability as much as by innovation.