To see our website at full performance please update your browser or use other modern browsers such as Chrome, Firefox or Opera.
 
 

To Barcelona and back again: Connected Car learnings from MWC

knight rider woo woo

It has been just over a month since Mobile World Congress 2017, where I was part of the team that demoed the world’s first in-car payment system. MOBGEN developed the solution in cooperation with Shell and Jaguar Land Rover, and I had the good fortune of being able to take part in the project from its inception to its production deployment in the UK just 2 weeks before MWC, and of course the actual preparations for the MWC demo itself. As it sometimes goes with demos, no amount of planning prepared us for the complexities we had to face at the venue in Barcelona, ranging from demo car configuration to 4G and WiFi networks congested to a level simply ironic for the name of the event.

After 3 days of scramble, our demo eventually came together and we ended up having an excellent MWC, including, but not limited to, winning a GloMo award – making a nice hood ornament for our demo vehicle:

IMG_20170228_162017

Sitting in my hotel room after the event, still sleepless from adrenaline, I took account of some of my learnings in the present of connected car related development, and contemplated on what the future might hold for us on four wheels.

Your car is your new screen

As your TV today is basically just another (big) output media for your cloud account (be it Apple TV, Netflix, or Youtube), so will be your fridge, your lamps… and your car. “Big multimedia” is coming, and is ready to disrupt this market too. The “car as the new screen” concept creates a whole new ecosystem and the industry is scrambling to be first and grab market share in providing you with the hardware and platform for it.

Samsung buying Harman, one of the big car multimedia solution suppliers for $8 billion is a clear sign. Because even though you might know Harman from your harman / kardon speakers and amplifier at home (or AKG; or JBL), 65% of their revenue is actually automotive related, as they supply some of the most high profile car makers with various hardware and software components.

Qualcomm, the company that’s probably manufacturing the brain of your phone, well… if it’s not a Samsung, announced their Connected Car Reference Platform for the same reason: hitching a ride in your car, but in a good way.

Microsoft integrating multiple mapping solutions (most recently: HERE Maps, the Google Maps alternative owned by BMW, Audi and Daimler) into their Bing Maps platform has a similar goal: to offer a viable reason to become an additional layer between the driver and the actual data.
Google and Apple are, of course, already well known contestants for a prime position in your car dashboard, with their Android Auto and CarPlay solutions. Amazon, a new but very strong contender, partnered with Ford to couple Alexa with Ford’s existing SYNC platform.

And naturally, the “old industry” doesn’t want to be left behind, so Valeo just got themselves a self driving vehicle permit in California. Intel just bought Mobileye, providers of autonomous vehicle and crowdsource mapping solutions. And Bosch just partnered with NVidia to create a self driving car computer. To add to this, we actually used Bosch technology for the in-car payment solution we demoed at MWC. This only scratches the surface, but they do give us a clear picture of what is happening in the automotive industry.

You need to treat your app’s security on a new level

What is also evident is that while the primary focus is on the multimedia in your car, everything is connected now: your multimedia experience is tightly integrated with the core telematics, safety and security systems of the vehicle. This has fun applications, like turning your car into Mario Kart controller.

Unfortunately it also has less fun side effects, like when hackers turn your car into their Mario Kart controller; or kill it remotely, while you’re driving on the highway. Or simply switch off your alarm from their phone, using a hacked remote control app.

When multimedia capabilities of cars covered CD changers and possibly Bluetooth phone connectivity, this wasn’t an issue. Today, when your phone app taps into the system that controls your alarm, it suddenly is. To blatantly steal Bruce Schneier’s phrase: computer security is now everything security. And everything includes cars.

Safety and sensors: vehicle and app are now much closer

Also, with platforms evolving, so is safety related legislation: the US Department of Transport now proposes a new rule that would mandate vehicle-to-vehicle communications, to ensure your connected car is also connected to the other connected car to avoid accidents.

On the positive side, with the new platforms bringing app and vehicle that much closer, we have a lot more exposure of the vehicle’s inner soul, via exposed sensor data. The industry is currently learning what can be done with this data. Basics include using data from the car’s much higher quality sensors in your app, like GPS or movement sensor; or differentiating app experience based on, for example, whether the car is moving or not.

There are more advanced concepts too, like combining sensor data with one of my other favourite things, gamification. Imagine your car sensors (while talking to other cars to keep the feds happy), also expose your good and less good driving behaviour to a gamification app that turns them into points and badges. You don’t have to go far – our Shell app already lets you collect badges for smooth driving.

There are new dimensions of complexity for app developers

An idea that was taken from your phone or tablet, with monthly security patches and system upgrades, car manufacturers are doing over-the-air updates to their cars now. They are learning what the mobile ecosystem has been learning for some time: how to deliver software updates, without screwing it up, at scale. This gives them the implicit opportunity to build their own ecosystems; when it comes to cars, manufacturers obviously control the platform itself.

For app developers wanting to integrate with this ecosystem, this brings three new dimensions of complexity.

First, as an app developer, you need to decide the core platform to go with. Are you building on top of an integrated platform and possibly putting some extra effort on your customer, or do you use tethered mode and let the phone app be the brain of the operation? It is a multilayered question in itself where there is no one good answer. You need to look at how your app will be used, how deeply does it have to be integrated, and how simple each option is for you as a developer, in terms of initial and maintenance effort (how well you know the platform, how stable you feel the platform to be, how widely used is the platform of your choice, and does that fit your desired target audience).

Second, you need to decide what car platforms you will integrate with. Think of it as the Android fragmentation problem, except instead of different underlying hardware to integrate your app with, you have a galore of underlying SDKs and APIs coming from different platforms that you want to integrate your services with – preferably still with a unified user experience at the end.

Then, once you’ve made your choices and integrated with a selection of connected car platforms, you will need to start planning your incremental upgrades to follow SDK upgrade cycles of platform providers, because remember? The car is your new screen (and speaker, and keyboard, etc.), and the driver to those peripherals just changed.

And don’t forget: all of these have to be tested.

You need to rethink the app testing

Any changes that you make in your connected car app, you need to test them. You need to test with them.
And depending on the level of integration you do, testing a simple change could bring an unexpected amount of effort when it comes to validating it on a car.

Similarly to car manufacturers learning the joys of OTA upgrades, app developers now have to learn that with connected car integrations, there’s stuff you simply cannot test in a test lab. This probably sounds trivial, seeing the amount of road testing going into autonomous vehicles. But while it is trivial to extensively test a self driving vehicle for safety reasons, you also want to test anything sensor integration related, for example, in a real life scenario.

An additional complexity is hardware fragmentation such as car hardware changes, and components getting incremental hardware upgrades. And while car manufacturers do a good job in keeping their platforms (SDKs, APIs, etc) consistent, mundane things like a change in the speed of an underlying hardware component can cause unexpected behaviour in your app.

All these bring additional complexity in testing even for smaller changes, and have to be reflected in test planning.

More than ever, UX design has to drive app experience

One of the main drivers in creating the in-app payment solution for Shell and Jaguar Land Rover was simplicity. We wanted to do whatever we had to do, with minimal screen taps and minimal driver distraction. When the Google Drive app on my phone provides me with an unexpectedly complicated user experience, well that’s sad, but at the end of the day, it’s just another couple of extra minutes from my life. When your in-car app drops one unexpected dialogue that potentially distracts the driver from driving, you might have your name associated with an accident in the press, like poor Mr. Potter who is now also remembered for having a supporting role in a Tesla crash.

JLRShellMotorist

As a result of our focus on simplicity, the number of possible taps you can make in our app whilst driving is exactly one. At the pump, when stationary, you can complete your refuel act in 3 easy steps.

We are still in the early stages

Last but not least, our project and our MWC demo has taught me that the journey has only just begun.
As the industry is in disrupted mode and everyone is scrambling to find their positions, so is the supporting industry exploring and learning what can be done with this new platform. My (obvious) guess is payments will always be at the forefront, as we’ve also been implementing solutions to pay for anything. We are also going into alternative controls, like how best to interact with an app in the car, extending, for example, on what Alexa is doing now.

We can expect closer integration with the other peripherals that users have (currently most importantly the smartphone, but this may change). We are in the early stages of understanding exactly how it is possible to make motorists lives easier.

As the underlying technology makes more and more things possible (including, but not limited to a significant increase in sheer bandwidth and processing power the cars provide), connected car platforms will evolve and will increase focus. Maybe some middle layers will grow stronger to address the fragmentation issue, and make the life of app developers easier.

And, naturally, the rise of autonomous vehicles will fundamentally change the way we use our cars, creating an entirely new segment with a different focus.

Whatever happens, we at MOBGEN and our MOBGEN:Lab will be continuing to happily tweak away with any improved or new technologies that the future brings!