http://www.toptal.com/
Whenever you
see a headline that poses a question, the answer to that particular question is
usually a big fat no. However, when it comes to software development for
next-generation car infotainment systems like Android Auto and Apple Carplay,
that’s not the case. If you can spare a few minutes, I will try to explain why.
We’ve all
heard talk of autonomous, driverless cars for years, and many of us had a
chance to try out some rudimentary implementations of such technologies, which
are slowly making their way to mainstream cars. This post won’t deal with those
for the following reasons: driverless cars are still years away, they will be
closed for development, and they will not create a new market for developers,
not unless you want your car to be parallel parked using code written by a
16-year-old coder freelancing for $5 an hour.
However, cars
with next-generation connected infotainment systems will create new
opportunities on several fronts. In case you already have an infotainment
system with a neat touchscreen and GPS in your car, please accept my
condolences; it’s about to become as obsolete as a Nokia 3310 compared to an
iPhone.
Next-gen infotainment platforms are to current
systems what smartphones are to feature phones.
I know that
is a bold statement, and many of you won’t agree with me, but I like to kick
off on a provocative note. I will do my best to change your mind, and if I
fail, feel free to let me know in the comment section.
Evolution Of
Car Infotainment Systems, Or Lack Of It
So what’s so
wrong with the current generation of car infotainment and navigation systems?
How come many of us chose not to buy them? Why don’t we see a lot of
development in this niche?
It all boils
down to a combination of technical and economic considerations. Consumer tech
is rendered obsolete in years, roughly two product cycles for smartphones,
three to four cycles for desktops and laptops. That usually translates into two
to five years. Naturally, as products mature, their lifecycle is extended as
well.
The car
industry does not work that way, so very few of us go out and buy a new car
nearly as often. In fact, many new cars ship with three- to five-year
warranties, so most people are unlikely to sell them for five or more years.
Cars are built to last a decade or more, and they can’t be upgraded like
desktop PCs, or receive OTA updates like our smartphones.
But hold on,
why don’t carmakers simply install off-the-shelf technology employed in tablets
and smartphones? Why do we still have expensive dials if it would be cheaper to
replace them with a high-resolution panel used in $200 tablets? The answer is
simple; it wouldn’t work.
Automotive
electronics are a world apart from consumer tech. While they can be based on
similar chip architectures and technologies, they need to be a lot more
durable. Unlike your iPad, your car infotainment system has to put up with a
very hostile environment and deal with loads of potential issues:
·
Constant
vibrations and G forces.
·
Extremely high
and low temperatures.
·
Ability to
stand up to high levels of humidity, or the occasional splash of water.
·
MTBF has to
be much longer.
·
When they
fail, they need to fail safe.
·
Infotainment
systems are integrated with numerous other components.
·
Legal and regulatory
issues must be addressed.
I could
expand this list, but I think it’s enough to prove my point; a car infotainment
system and an iPad don’t have that much in common. They may share the same DNA,
but the same goes for a MacBook Air and a Panasonic Toughbook.
The good news
is that developers needn’t be concerned by any of these issues, because they
will be addressed by carmakers and tech companies trying to slide their foot in
the door and grab a piece of this emerging market. That leaves us with good old
chips and operating systems, and whether they’re in a desktop, smartphone,
smart toaster or a new car, they all speak the same language; they all execute
code.
What Sort Of
Technology Is Coming To Our Cars?
A number of
tech heavyweights, including Apple, Google, Texas Instruments, and Nvidia, have
already entered this market. You can already buy cars equipped with some of
these systems, and some solutions like Nvidia’s Tegra-based infotainment
platforms, have been on the market for years.
The next
obvious step is to open up these platforms and get more brands, and consumers,
on the bandwagon.
Google Android Auto and Apple CarPlay are bound to dominate this space for years to come. In case you
would like to check out a head-to-head comparison of the systems, you can headover to CNET.
Now, these
platforms don’t have much in common with driverless cars, autonomous cars or
whatever you choose to call them. They are merely a replacement for the “dumb”
infotainment systems we have today. Here is an analogy that should explain
what’s about to happen: next-gen infotainment platforms are to current systems
what smartphones are to feature phones. They have a lot more potential for
future development, integration with other devices, fast mobile broadband
connectivity, and so on.
In terms of
hardware, we will see more powerful System-on-Chip (SoC) solutions, capable of
providing fast 4G data, stunning graphics, improved GPS capabilities, and even
some cutting-edge features such as motion tracking.
For example, Nvidia is trying to leverage its GPU technology to enable motion tracking
that should provide drivers with better situational awareness. This does not
mean we’ll end up with driverless cars powered by Nvidia SoCs, but the
technology could be used to look out for obstacles when parking, cars in our
blind spots and so on. Not long ago, the computational power needed to pull
this off was reserved for professional graphics solutions, but the latest crop
of Nvidia Tegra processors features 192 GPU cores, or CUDA cores to be precise.
Upcoming Tegras will feature even more powerful CPUs and additional CUDA cores
(256 and more cores).
Even the current generation is powerful enough to enable the development
of autonomous cars, let alone vehicles with some rudimentary motion tracking
features. In case you are interested in the geeky details, you can checkout this Nvidia blog, detailing
how a Jetson TK1 development board can be used for low-power sensing and
autonomy.
The really
good news is that the industry will be able to use vast amounts of CUDA code,
developed for discrete graphics cards. It will work on Nvidia mobile platforms
as well. The bad news is that Android Auto and Apple CarPlay simply won’t
harness this potential, at least not yet. Instead, they will act as “second
screens” for our mobile devices.
Bottom line;
hardware will not be an issue.
No comments:
Post a Comment