How Ford has slammed the door on Silicon Valley's autonomous vehicles drive
Owning the smartphone-dashboard interface
Detroit and Silicon Valley aren't just 2,000 miles apart – they're on different planets, culturally speaking. One is the home of America's automotive industry, a heavily regulated, ultra-conservative sector focusing on high-volume, low-margin sales. The other houses companies that deal in high-margin information and digital services, acting first and begging forgiveness later. They are also in competition to own what some are calling the next personal computing platform: the car.
The recent focus is less the embedded systems that run vehicles – Linux won that battle – and more the data connections to deliver so-called "infotainment" to those inside and beam diagnostic data back to the manufacturer.
Gartner reckons on a five-fold increase in such units globally by 2020 to 61 million.
With those units come opportunities. Now, the sale isn't just about the car; it's about the subscription revenue that you can garner from apps and online services. It's going to boost aftermarket revenue opportunities in a chronically low-margin industry. The car is the next PC platform.
No wonder Google, Apple, and Uber's driverless truck startup Otto, to name just three, are pushing hard on autonomous vehicles.
And no wonder car makers rallied together and fired back with the SmartDeviceLink Consortium (SDLC), for connecting smartphone apps to vehicles.
SmartDeviceLink (SDL) hails from Ford, based on its proprietary AppLink APIs for smartphones to talk to car systems that it open-sourced and contributed to the GENIVI Alliance in 2013 under the BSD license. GENIVI "OEM" members are listed as most of the major car-manufacturing brands.
Four years later Ford with Toyota have formed a consortium to promote the SDL as a way to connect smartphone apps to vehicle's smart dashboards (the head unit) and control them with the vehicle's human machine interface. First members are Mazda Motor Corporation, PSA Group, Fuji Heavy Industries (now the Subaru Corporation) and Suzuki Motor Corporation.
Why? Imagine manipulating Spotify via your steering wheel controls, for example.
But such services and the data they use or generate are only part of the story. Devices and data are nothing unless you also have developers – individuals and companies who build applications that can deliver the applications and services people need, or don't know they want, and that generate or gather data for the car manufacturer once the vehicle is on the road.
And therein lies the problem. If the world of enterprise has taught us anything, it's that success comes from attracting developers and developers attracted in sufficient numbers by the ubiquity of platform – by the fact that an operating platform can straddle different hardware. Software that lets you – as close as possible, if not entirely – write your app once so it runs anywhere without rewrites.
Divided you'll fall
That's where the car giants have failed. Ford has AppLink, but GM has its own developer framework called MyLink. For the car to be the new PC, there must be a standard for it so we're not developing the same app for the four-wheeled equivalents of the pre-PC Osborne, Kaypro, and Apple II.
James Hodgson, industry analyst at ABI Research told The Reg: "The way OEMs have approached adding functions to their head-end entertainment units is the way that the old Nokia would have done. What they want was baked in from the start. The key to maintaining relevance in the vehicle is to have this application framework and bring in developers who are innovative."
"Fragmentation is certainly hindering the uptake of in-vehicle subscription services. However, music subscription services aside, there is certainly a potential for applications in the future," says Sam Barker, research analyst at Juniper.
"For these services to be successful, partnerships that lower the level of fragmentation in the market will be critical." But that's an area where Silicon Valley has history in building the kind of hardware-straddling ubiquity that attracts developers.
Apple has been trying to shift from a maker of expensive computers and laptops into a service firm, with iTunes, the iPhone and iPad. In March 2014, Apple announced the CarPlay system, released in 2014. Google has Android Auto, which it introduced a year later. Both are UI projection systems for smartphones, casting driver-friendly versions of the phone's display to the dash display.
Many car makers actually support these systems – including Ford, which supports both interchangeably.
No, you wouldn't swap mid-drive, explains Jeffrey Hannah, director for North America at car analyst SBD, but between drives. Bob might prefer to control the manufacturer's curated selection of supported smartphone apps via SDL. This doesn't stop his wife Alice from locking herself in Apple's gilded cage on the way to her yoga class, or their daughter Jane subsequently beaming all her personal data to the GooglePlex via Android Auto when she takes over.
The official line from car makers is that anything enhancing the driver's experience and giving them more choice is good. Informed by his own consumer data, Hannah buys this argument.
"They know that consumers want to safely use their smartphone in the car in the best way possible. The car makers made a compromise to say that the consumer is boss," he says.
So, SDL – in theory – creates one common framework for all car makers, making it easier for an app and service provider to get to your dashboard – no rewrites or ports. Smartphone apps are already in modern vehicles, though, with many car manufacturers already supporting one or both in-car options from smartphone vendors. These smartphone operating systems integrate their native apps and virtual assistants with your car. You can control them with your car's hardware.
So why bother with SDL at all?
The driving data goldmine
The answer is that there's something you can't easily access with your smartphone alone. Saying "OK Google, where's the nearest gas station?" works just fine, but "Siri, how far can I make it before I run out of gas" won't. Neither will asking your voice assistant whether the tires need inflating, or how safely you're driving. This kind of thing needs telematics data from the car, sent from different sensors along the vehicle's local network, called the controller area network (CAN) bus.
"This data is the gold," says Hannah. It includes GPS, fuel, acceleration and braking data, along with diagnostic info; anything, really, that a sensor in the car can send.
It's also data that the car makers are keeping for themselves. Apple and Google may have a place in many dashboards, but there is a Chinese wall between their smartphone platforms and that CAN bus data.
Car makers typically put that down to customer privacy and security. It's difficult to control that info when it's on the smartphone platform. Most customers aren't savvy enough to work out their privacy settings when installing a dodgy knock-off version of Flappy Bird, let alone protect their telematics data from snoopers.
Toyota, one of SDLC's founding members and the planet's second largest car maker by sales, plans to start offering its own telematics system build on SDL "around" 2018, the firm said.
There's also a glaring economic and political issue here.
"[Car makers] haven't slammed the door on Android Auto and CarPlay but they do want their own parallel platform where there would be certain apps that are only accessible on SD Link [SDL]," says Roger Lanctot, director of automotive connected mobility in the global automotive practice at market analysis firm Strategy Analytics.
"The challenge for Ford is to enable the discovery of those apps that are only available on SD Link and build customer retention proposition around those applications."
Bloke to insurance firm: How's my driving?
If developers are critical for success, what kinds of applications are we talking about?
Right now, a lot of Sync apps focus on navigation, news, ecommerce and music, which are all things that phones can do easily enough. Differentiation is coming, though. Or so SDLC reckons.
"There are lots of vehicle-centric downstream opportunities where at this point Android and Apple may not have the best apps at all," says Doug VanDagens, Ford's global director of connected vehicles and services, and an SDLC board member. "Where we're going next are that there are arenas in repair, insurance, and diagnostic information. There are other players that want to take advantage of what the vehicle offers."
He gives examples such as crash analysis, where sensors might be used to better describe what happened in a collision to insurance companies.
Ford has already announced a partnership with IVOX, developer of the DriverScore app that tells insurance companies how you're driving and potentially lowers your premiums.
Coming this spring, the app connects via AppLink and slurps vehicle performance data with the driver's permission. IVOX then stores the data, aggregating it to produce a driver's score. It sends that score, rather than your precious vehicle data, to insurance firms who will then adjust the premium to suit both grannies and boy racers. At a 2017 CES hackathon in January, Ford also demonstrated a not-yet commercially deployed app built on IBM's Watson. It accessed vehicle fuel data to tell particularly dim drivers that fuel was low, and then alert them to gas stations en route.
More impressively, at CES it also announced integration with Amazon's Alexa, enabling users to control functions directly via voice. You can start your vehicle and warm it up from inside the house on a cold day. Those people unlucky enough to park it unlocked in darkest Romford can quickly ask Alexa to remediate the situation from afar. Just don't say anything that sounds like "dollhouse".
What's next? What else might be coming down the pipe? BMW's interview here was tantalising. Its execs talked about a world where cars ship with the appropriate hardware features – everything from heated seats to extra horsepower – that were then turned on by software. Bum getting a bit chilly over the winter? Upgrade to heated seats. Going on a hilly road trip? Rent some extra horsepower.
We're already seeing Teslas ship with features that get turned on later, so this isn't inconceivable. Not that Musk will probably want anything to do with SDL, though – that firm is going its own way.
The big question: Will SDL gain traction?
iOS saw off a lot of competition. Not just from Microsoft's Windows Phone, which some analysts reckoned was destined for success. The open-source camp churned over the years with the merger of the Linux Phone Standards (LiPS) Forum and the Linux Mobile Foundation (LiMo) that was rebranded the Tizen Assocation. There was also the Linux-Foundation-backed MeeGo that the Foundation dropped for Tizen. Tizen's members include Motorola, NEC and Samsung but despite this commendable moving of the chairs, it's Android that dominates open-source handsets.
Car makers are signed up to SDL: Mazda, PSA Group (Peugeot, Citroën), Fuji (Subaru) and Suzuki. On the app side there's parking app ParkU with Honeywell promising to let you turn your home thermostat on by barking at your steering wheel. Interestingly, Google-owned Waze has also signed up, which suggests that things are – after all – affable between car makers and smartphone OS vendors.
And yet, as Apple's success has proved, numbers of founding members are no indicator of success.
Still, Juniper's Barker is sceptical about the initiative. "SDL has entered the market as an open-source alternative. However, it is unlikely to be enough of a differentiator to compete with services like CarPlay and Android Auto in the long run," he says. "These two systems will increase their in-vehicle offerings through the introduction of new apps themselves."
It's early days for SDL, but also for the smartphone OS vendors. With autonomous driving likely to change everything in the next few years, we're so early into the journey that we haven't even turned the ignition yet. One thing's for sure: cars in 15 years are going to be far different than this year's model. Maybe there won't even be a steering wheel to poke at that point. ®