BCG-WEF Project: DRIVE-A: Autonomous Vehicles

Advancing Industry Collaboration in Vehicle Software

Most of us are used to judging automobiles by their performance, muscle, handling, and ride. For nearly a century, those attributes were the target of virtually every improvement that automakers chased. Those days are decidedly over. As we drive toward the 2030s, the role of software, not hardware, in determining vehicles’ innovative features and capabilities will be a primary focus for the automotive industry and ultimately shape how drivers feel about the cars they own. Indeed, the new generation of cars are commonly referred to as software-defined vehicles (SDVs).

SDVs are at the intersection of three powerful trends: the emergence of electric cars, the increasing importance of smart digital experiences inside automobiles, and the evolution of vehicle autonomy. In a fully programmable SDV, digital components—responsible for safety, comfort, infotainment, drive train, and performance—would be controlled by software and regularly deployed and updated through over-the-air networks.

But although the future is certainly bright for SDVs—BCG estimates that the market could nearly double between now and 2030 to $660 billion—one very large danger sign could dim their prospects: the glaring need for collaboration across the auto industry and technology companies to advance the transition to the software-defined vehicle.

This is especially true when it comes to one of the most crucial elements of an SDV: the vehicle software platform, the control center of the vehicle consisting of operating system and middleware. This end-to-end digital architecture links all the components and code needed to manage a modern-day automobile. (It should not be confused with the vehicle platform, which is mechanical in nature, taking signal input from sensors and wiring to initiate an action or response from the vehicle.) The vehicle software platform requires coordination among multiple stakeholders to address interconnections of components and operating systems, decoupling of software and hardware (a prerequisite for its scaled development), and regulatory standards that vary across geographies.

Multiple industry consortia, each with its own objectives and membership composition, are working to improve collaboration, but a unified direction has not yet emerged. In an attempt to create a blueprint for collaboration, BCG and the World Economic Forum identified a series of priority areas for cooperation. The first includes the key components of the vehicle software platform, encompassing rules and protocols, app development, and technology integration, among other things. The final area involves coordination and management of these software initiatives to ensure aligned objectives and deliverables.

To implement this blueprint successfully, there must be clarity in the roles and responsibilities of the various consortia as well as a unified direction that maximizes synergies among existing efforts.

Advancing a Safe Vehicle Autonomy Future

Of the three trends propelling the development of software-defined vehicles (SDV)—electrification, smart interior digital experiences, and autonomous driving—the last may represent the most radical and far-reaching departure. Fully self-driving vehicles will open up new mobility opportunities for seniors, shut-ins, and people with disabilities, while offering hands-free time to drivers.

And, perhaps most important, vehicle autonomy could result in considerable safety benefits. Currently, 94% of vehicle accidents are due to human error; many of them could be prevented by self-driving cars. Although full vehicle autonomy is a long way off, the partial autonomy capabilities already available can contribute to road safety and an improved driving experience. That is, if certain critical obstacles involving the shared responsibilities of the driver and the vehicle are overcome.

The Society of Automotive Engineers outlines six levels of autonomy, ranging from 0 (fully manual) to 5 (fully autonomous). The most advanced autonomous features in commercially available vehicles can be categorized as high Level 2 (partial hands-off driving, automated lane changing, and traffic light detection, among other things) to Level 3 (the system performs all driving tasks if the right conditions apply but will ask drivers to step in when necessary). However, an analysis by BCG and the World Economic Forum found that these self-driving features have not caught on sufficiently with drivers—even among those who own cars with autonomous capabilities—or have deficiencies that hinder notable safety gains.

The limitations are, in part, the result of unfamiliarity: fewer than 5% of consumers have an accurate understanding of L2+/L3 capabilities and risks. In turn, the perception of value for semi-autonomous features and willingness to buy and use have been depressed. Beyond lack of knowledge, operational shortcomings are also a problem. Nearly half of users in a BCG survey said that they feel they must pay more attention to their vehicles when a semi-autonomous system is engaged. Indeed, a big challenge is ensuring that the driver is alert enough to re-engage in handling the vehicle when necessary and that the takeover process is simple. Many drivers are worried that if they allow their concentration to flag, they will be caught flatfooted in a dangerous situation.

To overcome these obstacles, autonomous vehicle players must agree on the nomenclature for the various levels of self-driving as part of a larger effort to educate drivers and set realistic expectations about the benefits of and evolution in this sea change in mobility. At the same time, companies must rapidly improve autonomous features so that the concept of autonomous vehicles is not as far-fetched as it may seem to many car owners today.