The key to self-driving cars is software that can interpret all of a vehicles’ sensors and learn to mimic the driving skills and experiences of the very best drivers. Google is the current technology leader in this arena, according to a report from industry analysts IHS Automotive: “Google Self-Driving Car Strategy and Implications”. IHS estimates suggest Google has invested nearly $60 million so far in autonomous vehicle research and development, at a run rate of nearly $30 million per year.
Unlike traditional vehicle manufacturers, Google also has the ability to leverage adjacent technologies and learnings from its other projects and investments—including robotics, drones and related technologies that help automotive operations, such as neural networks, artificial intelligence (AI), machine learning and machine vision. This provides Google researchers additional expertise not available directly to traditional OEMs.
No other company has as much relevant technology to advance autonomous driving software.—Egil Juliussen, PhD., senior research director at IHS Automotive and author of the report
Toyota’s early November announcement of a $1 billion, five-year investment in AI, driverless cars and robotics is likely partly due to Google’s rapid technology advances, IHS suggests. (Earlier post.)
Two primary research and development strategies to achieve self-driving vehicles are in place today: evolutionary and revolutionary. Most traditional automotive manufacturers are on the evolutionary track with their R&D efforts, continuing along the current path of improving advanced driver assistance systems (ADAS) to partial self-driving, and eventual full self-driving vehicles.
On the other hand, Google leads the revolutionary approach and will have a major impact in the coming years, according to IHS. Uber is also beginning to implement some of its own R&D in this arena, as it works toward solutions for the next stage of its disruptive transportation strategy.
Google’s self-driving car software is already performing better than nearly all drivers in the vast majority of traditional driving situations—at least in good weather, according to IHS analysis. However, as it continues its development, Google still must discover and teach its software the “once in a million” events—such as performing under diverse weather conditions, unique roadwork, specific traffic situations and other nontraditional driving situations.
Google is in a unique position to leverage adjacent technologies for developing self-driving car software, and its strategy and goal is to provide the software and map infrastructure to allow mobility services to anyone—via fleets of driverless cars—within a decade or less.—Egil Juliussen
As a result, the car-as-a-service approach may see significant opportunity. Currently, these services require drivers. In the future, it is exceedingly likely that these and similar services may be able to operate with a driverless approach.
Car-as-a-Service (CaaS). Car-as-a-service (CaaS) opportunities are becoming a new driving force for urban transportation. CaaS is essentially an extension of car-sharing, but via driverless vehicles. Google’s strategy is to provide the technology infrastructure, maps and software to make CaaS happen sometime after 2020. IHS Automotive analysts believe Google is currently developing the software and maps that can be the basis for driverless vehicles in five years or so.
IHS estimates that the deployment of driverless CaaS will begin before 2025 and will have increasing impact as technology advances and driverless car volumes grow. In addition, autonomous driving and the associated costs will dramatically lower the cost of mobility services as a whole.
CaaS also will provide car mobility services for anyone, since no driver’s license is needed and it will be an affordable transportation solution for a large portion of the global population. IHS Automotive estimates there are currently 6.2 billion people globally without a driver’s license, or nearly 85 percent of the world’s population. Driverless cars can also provide mobility services for packages and other goods requiring transportation.
IHS Automotive currently forecasts that nearly 12 million self-driving and driverless cars will be sold globally in 2035, which will reflect about 10% of total global light vehicle sales. With continued increasing R&D activities in autonomous driving technologies, the current forecast could be too low.
EVs well-positioned for autonomous enablement for CaaS transportation. The driverless cars, and future mobility services based on them, will also provide a major opportunity for electric vehicles, IHS suggests. Driverless car mobility services will mostly happen in urban areas and will primarily be short trips.
These characteristics favor EVs as the powertrain for driverless cars; they can easily re-charge themselves using existing and growing public charging networks as needed between trips, which eliminates any range anxiety. IHS Automotive forecasts that global EV charging stations will grow from 650,000 in 2015 to more than seven million in 2021—excluding home charging outlets.
Mega-cities and other large urban areas in the future also will prefer the low emissions of EVs and as a result, should be keen to implement fleets of driverless EVs in their communities. Added benefits of EVs as part of CaaS may also lower congestion and help to resolve other traffic-related challenges.
IHS Automotive forecasts that the global production of battery EVs will grow from 273,000 in 2015 to 1.3 million in 2022. Global production of plug-in hybrid EVs is projected to grow from 179,000 in 2015 to over 2.4 million in 2022, according to IHS estimates. Driverless car fleets for CaaS are likely to greatly increase the sales of EVs after 2025.