Intel’s Whole-Vehicle Approach to Boost Automakers Profits

Estimated reading time: 4 minutes

Nearly every aspect of the automotive industry is undergoing significant transformation – from business models and supply chains to exciting new in-vehicle experiences, including artificial intelligence. Unfortunately for automakers, these changes often come with increased costs, especially considering the expense of high-performance AI-enabled system-on-chips (SoCs) and, for electric vehicles (EVs), ever-larger and more expensive batteries.

The industry must reduce costs. Yet with the unrelenting march of technology advancement, how can they do so at a profit while delivering the next-generation experiences that consumers desire?

The answer is to adopt a systems-based approach.

Adopting a Holistic Strategy

The myriad of cost challenges can’t be solved one socket at a time. Instead, Intel proposes a holistic, system-level strategy.

Rather than simply building better individual components (which we do), we connect the dots to deliver systemwide benefits. We focus on integrating three key elements: software defined in-cabin compute, intelligent energy management and data-center-like workload management. The combination of these advancements delivers a multiplier effect over trying to cost-optimize any one aspect of the vehicle.

The Intel approach looks at the vehicle systems as a “whole” and allows for the seamless movement of workloads between software-defined central compute systems and software-defined zonal compute subsystems, ensuring maximum flexibility, optimal cost and performance with significant energy efficiency benefits.

Breaking Down the Silos

Current vehicle architectures are siloed, leading to inefficiencies. For example, many EVs – even when off – support a feature that still monitors the external cameras for security threats or to recognize the driver as they approach. Typically, this feature is supported on the vehicle’s in-cabin compute subsystem, which due to its high-power consumption, puts unnecessary drain on the battery even when the vehicle is off.

This workload doesn’t have to stay resident on the software-defined central compute system. If we use Intel’s software-defined zonal controllers to handle camera streams, we can embrace data center application orchestration concepts and migrate the workload to a lower-power device (in this case the zonal controller) and wake the central compute system only when needed. It would save energy, improve efficiency and reduce the total number of electronic control units (ECUs) in the vehicle by consolidating workloads dynamically onto a software-defined zonal controller.

Moreover, integrating intelligent power policies with control systems can reduce energy use across the vehicle. For instance, turning off the ADAS ECU while the vehicle is charging or adjusting vehicle power utilization based on environmental conditions can conserve energy significantly. In winter in Detroit, turn off the A/C ECU. In summer in Phoenix, turn off the seat heater and windshield wiper ECU.

These may be simple examples, but they offer a profound change in what a system-level view can do to a vehicle’s architecture.

Apply this concept across the entire vehicle with every ECU controlled from a centralized power management controller and there are infinite possibilities for conserving energy. That will make every vehicle more efficient, regardless of whether it’s an internal combustion engine (ICE) vehicle or an EV.

These strategies aren’t new. They have transformed the PC industry, leading to longer battery life through standards like the advanced configuration and power interface (ACPI) specification that allows all power-consuming devices on a PC platform to be discovered and deterministically controlled. This is, in large part, how the PC industry transformed from early laptops that would hardly last an hour to the all-day battery life we enjoy today. And this thinking is already being translated into the automotive industry with the new SAE Vehicle Platform Power Management Standard (J3311), which aims to apply these proven PC concepts to vehicles.

An Integrated Architectural Philosophy with Roots in the Data Center

Software-defined design done right is an architectural mindset – a philosophy that says compute, memory and I/O are pooled and shared resources that can be dynamically allocated with freedom from interference – to whatever workload is at hand. Taking such an approach changes the way we look at a vehicle’s electrical/electronic (E/E) architecture from fixed-function sockets with a 1:1 mapping of application to silicon, to a pool of resources that span multiple sockets and enable new system-level approaches to deliver the experiences consumers demand.

In short, it’s a data center, not a phone/tablet approach. And Intel has done this many times before, making the company perfectly suited to help the automotive industry through this pivotal transition.

The Wheels of Change are Turning

Transitioning to software-defined, sustainable and scalable vehicles isn't easy. But it’s going to be even harder if automakers try to evolve the vehicle architecture one socket at a time.

Adopting a holistic system-level view, with the right silicon and features designed in coordination, will open new pathways to profitability, an approach Intel is uniquely positioned to lead.