We fall in love with cars because of their hardware: their sleek frames, the growls of their engines, the luxurious amenities within them. But today software, not hardware, truly delivers on the demands of consumers. Software drives the electronic control units (ECUs) that power everything from dashboard instruments to safety features to powertrain components to in-vehicle infotainment systems.
However, the success of ECUs in delivering state-of-the-art functionality has created a new challenge. The number of these devices in the average car has doubled in the past 10 years; many cars now incorporate more than 125 ECUs. These units take up increasing amounts of space. They draw power, and their weight decreases energy efficiency.
ECU consolidation is an obvious answer, but how? How do you deal with the increasing software complexity? How do you consolidate without an explosion of integration issues? How do you avoid a massive new testing effort that could jeopardize time to market? What can we learn from other industries?
This article summarizes a new approach to ECU consolidation that can tame the complexities with minimal impact to the testing process. This approach can actually accelerate innovation in ECU-based functionality, and it can drive new competitive advantages for forward-looking automakers.
The challenges of ECU consolidation
The sheer number of ECUs in today's cars is not the only complication for consolidation. The functions performed by ECUs are also becoming more sophisticated and complex. New features and capabilities such as adaptive cruise control, digital instrument clusters, and car-to-car or car-to-fleet communication are great for consumers, but they add complexity to an already difficult software development and testing process.
Each ECU also has its own infrastructure components (such as power supply, bus, and diagnostics) and its own requirements for functionality, safety, security, and dynamic behavior, requiring the use of multiple platforms and toolsets for development and testing. In addition, ECUs may have differing Automotive Safety Integrity Levels and ISO 26262 requirements.
Selecting among the array of proprietary and open-source development, testing, and management options further complicates matters. How do you test and debug with all these permutations and complex interactions and interdependencies? How do you meet certification requirements in this environment?
Above and beyond the technical intricacies of ECU consolidation, there are business issues to consider. ECUs are typically purchased as separate items, potentially from separate Tier 1 suppliers. If ECUs are consolidated and combined, who builds them? Who manages the multi-vendor sourcing, integration, and licensing issues? Who is responsible for maintaining and upgrading their functionality? How are faults isolated to an individual function?
No single technology partner can resolve the business challenges of ECU consolidation. Wind River's perspective is that a technical solution to consolidation will accelerate the resolution of the business challenges. Therefore, we have dedicated our resources to that effort.
We can certainly learn from other markets that have solved the technical and business challenges. A great example is the aerospace industry, where integrated avionic systems have been launched successfully. The technical solution of consolidation -- and the clear definition of roles such as platform provider, application provider, and software integrator -- have helped overcome the business challenges. With its long record of automotive experience, Wind River is in a position to combine the expertise of these markets.
Conventional options for ECU consolidation
Several approaches to ECU consolidation have been attempted, with mixed success. The most obvious approach (shown on the next page) focuses on software integration. Multiple applications run on a single ECU using a common operating system, such as OSEK or AutoSAR, with communication via a CAN or other standard bus.