From Software Enhanced to Software Defined: Developing the Next-Gen Vehicle

The mix of software and vehicles is hardly a new phenomenon. Embedded software has been a key enabler of improved emissions, fuel efficiency, and safety as well as new features since the 1970s. But the key word is embedded, which has resulted in software-enhanced vehicles. One of the big trends in the past few years and going forward is the software-defined vehicle (SDV), which dramatically expands the scope of what has been done before. 

The Expanded Scope of SDVs

With embedded software systems, the software code tends to be tightly coupled with the hardware platform it runs on. Until recently, that hardware was generally specified for precisely the functionality the vehicle was intended to be sold with. Embedded systems were also widely distributed across the vehicle and developed in a siloed environment with minimal consideration for shared computing resources. 

With the modern SDV, the architectural approach is fundamentally different. The software is generally being abstracted away from the hardware platform, providing both the hardware and software engineering teams more flexibility and enhancing design efficiency. 

It’s analogous to the evolution of mobile phones over the past 20 years. In the early 2000s, most phones were also highly software-embedded devices with little or no capability to add new functionality over the device’s lifespan. What little could be done was limited to a single device. When a new model came along, everything had to be done again. The emergence of the smartphone, starting with the Palm Treo and really taking off with the Apple iPhone and then Google Android-powered phones, changed the game. Each of these featured a platform that developers could use to access common services such as the phone, data modem, GPS, and camera. The manufacturers could update the underlying hardware from year to year while apps continued to function over multiple generations. 

With automotive embedded systems, the overall system tended to be brittle when features shared data. When the team behind one function made a change to code, it could often break other systems. In the SDV, application programming interfaces (APIs) provide a consistent pathway to services and data. The underlying implementation can evolve and update along with hardware while applications can develop independently. 

This can be executed on a hardware platform that still resembles the distributed computing architecture of old, but increasingly manufacturers are making the transition to more powerful compute devices and beginning to consolidate that hardware. Building core hardware functionality including sensors and actuators into all variants of a vehicle rather than limiting it to specific trim levels reduces configuration management. 

With simpler but more comprehensive hardware, creating the SDV platform that can allow customers to select the functionality they want and even when they want it becomes possible. It also becomes possible to improve and increase the capabilities of vehicles over their lifetime, as Tesla has demonstrated over the past decade. This capability is a key factor for automakers that hope to double revenue in the next decade through subscriptions and services that could be delivered by over-the-air updates. 

SDVs Require a New Way of Working

Executing on all of this requires a new way of work, with new tools and processes as well as a different organization. Join Roger Ordman, Executive Vice President of Marketing at Aurora Labs; Scott Miller, Vice President of Software-Defined Vehicle and Operating System at General Motors; and Ali Kani, Vice President of Automotive from NVIDIA, for a conversation on the development of SDVs and vehicle software intelligence tools in a free Guidehouse Insights webinar on April 12, 2022. Register for the webinar and read the white paper on vehicle software intelligence.