Consider for a moment what the creation of a truly connected and intelligent vehicle could mean. First and foremost, increased safety. Also, improved fuel efficiency. Less congestion. Decreased pollution.
Even better? A driving experience enhanced with Intel Internet of Things (IoT) that can help make your commute safer, more productive, and enjoyable.
With the rapid convergence of commute and compute, the automotive industry is undergoing a radical transition not seen since the creation of the assembly line. And that’s presenting automotive and IT professionals with a unique opportunity: to take a fundamentally different approach to vehicle function, design, and construction.
Today, our cars are transportation machines that contain computers. Tomorrow, they’ll be intelligent data devices that actively help us take advantage of every moment on the road. Even the materials we use to build them may change; we could have plastic vehicles, or even biodegradable ones.
Sound too sci-fi? It’s not. Look at what we’re already doing with advanced driver assistance systems (ADAS). Twenty years ago, adaptive cruise control, automatic parking and braking, blind spot detection, and collision avoidance—and self-driving cars—were just engineers’ dreams.
Navigating a connected environment with V2x
As automakers race to build the ultimate ADAS, two camps are emerging.
The first is focusing on vehicle message passing, in which the car communicates with:
- Other vehicles (V2V)
- Infrastructure points, like signs and traffic lights (V2I)
- Bicyclists and pedestrians, via smart-phone apps and wearables (V2P)
- The cloud (V2C)
Combined, we call this “V2x.”
The main challenge with V2x is what to do with the incoming data. Many proponents feel that the best approach is to compile and present it to the driver to interpret and act upon.
But then there’s the danger of information overload. We know from managing cybersecurity breaches that it’s risky to treat all incoming threats equally. They need to be prioritized, so that people are alerted only to the most pressing concerns.
Plus, much industry work is required to ensure that V2x communication protocols are standardized. To get a serious systemic payoff from the technology, all vehicles must speak a common language. Disparate formats will both diffuse the safety benefits and confuse consumers.
Finally, V2x by itself is dependent on robust network connectivity. Not all wireless communication protocols are created equally in ensuring secured, deterministic message passing. And not all can achieve global ubiquity, efficiency, and cost-effective economies of scale.
Putting the driver in the passenger seat with onboard intelligence
The alternate camp says the way to increase driving safety, enjoyment, and productivity is to remove critical tasks from the driver and assign them to an onboard computer. To, essentially, create an autonomous car.
That means the individual vehicle reads and reacts to its surroundings, so it’s less dependent on connectivity and incoming messages.
There are challenges to address here, as well.
Security is as much a concern with autonomous cars as it is with V2x. You don’t want your connectivity or onboard systems vulnerable to hacking. And the profusion of discrete electronic control modules (ECMs) aboard an intelligent car presents a tempting array of attack vectors. Particularly since many are sourced from third-party suppliers and may go unchanged for years.
Then there’s the challenge of getting public buy-in. In the United States in particular, we have a driving culture that values individual driver control. If you ask your average driver if they want a computer to take over in an emergency, many will say no.
Joining the best of both worlds
So, my question is, why do we have to choose one or the other of those approaches? Why not blend the two to create an incredibly intelligent vehicle with an ADAS that can interpret and respond to its surroundings, and augment it with connectivity to cameras, LiDAR (Light Detection and Ranging), and radar? A vehicle that complements and enhances human cognitive abilities, rather than replacing them?
Reimagining the car, from the wheels up
To build this next-generation intelligent car, we’ll have to enable sensor output fusion and shift to a centralized computer—one that can mimic a human brain. And the vehicle will need 360-degree awareness.
This will require an unprecedented amount of compute horsepower in a small thermal envelope, scalable from vehicle to vehicle and generation to generation. It will also require the greatest ratio of performance per watt per cost per functional safety. Plus—and this is critical—the system has to fail safely and fail operationally.
So, as you can see, all of this calls for a major reimagining of car architecture. As well as unprecedented levels of collaboration between industries, government agencies, consumers, and special interest groups.
Developing the car as a platform for innovation
That’s why my team at Intel is made up of a combination of forward thinkers from the automotive industry, industrial control, functional safety, and overall platform design. As well as hardcore IT silicon specialists who have dedicated their careers to building mission-critical servers.
Together, we’re looking at the car as a holistic platform. We’re exploring the best and most efficient way to capture data, then securely move it from the edge to the central brain in real time.
No more siloed systems. No more just bolting on another box when you want to add an ADAS feature. We’re building a whole new backbone. One with a brain that assists as needed and controls when control is handed over to it. Essentially, we’re creating an entire car that’s an ADAS.
To do that, we’re borrowing best practices from the avionics industry. Think about how commercial aviation has changed. And how it hasn’t.
We now have onboard computers, autopilot, and collision avoidance systems, and they all make air travel far safer and more reliable. But we haven’t made pilots obsolete; they’re still the apex decision makers. And that’s how we think the next-generation intelligent car should operate.
Driving the evolution of automotive with Intel IoT
For more than a century, motor vehicle research and development has concentrated on engineering mechanical systems. Automatic transmissions. Air conditioning. More efficient engines.
But that emphasis is shifting as the worlds of automotive manufacturing and computing converge. In the future, the majority of vehicle research will focus on software, not hardware.
Intel IoT is uniquely situated to play a central role in developing a connected car; one that not only leverages more powerful and efficient computing, but can also help to enhance safety and security. We’ve been involved in mission-critical systems for decades. And we want to catalyze the development of best practices, standards, and platforms that will be used across the automotive industry.
That’s why we created the Automotive Security Review Board (ASRB) to convene top thinkers from across the technology and automotive sectors. The ASRB is exploring how drivers use cars, what they value, and how we can expand the user experience. And, above all, how to increase safety and reduce accidents.
Intel is also inviting industry experts to comment on our new white paper, Automotive Security Best Practices: Recommendations for Security and Privacy in the Era of the Next-Generation Car. We’ll publish revisions based upon feedback and findings from the ASRB.
Dreaming of the next-next-generation auto with Intel IoT
So today Intel IoT is focusing on building a vehicle with a central brain. What comes after that?
Well, we’re already thinking about it: A vehicle with a cognitive central brain.
Now that’s next generation.
To learn more about what Intel is doing to move the automotive industry forward, visit intel.com/automotive. To learn more about Intel IoT developments, subscribe to our RSS feed for email notifications of blog updates, or visit intel.com/IoT, LinkedIn, Facebook, and Twitter.