Your average car is a lot more code-driven than you think
FOSTER CITY, Calif.— The growing interest in smart and connected cars is getting tangible, both for car buyers and car makers, just as they get more complicated — a lot more complicated.
In fact, to really appreciate both the value of the car-related technologies that are being offered now, as well as what’s possible in the not-to-distant future, it helps to have a basic understanding of how modern cars function.
At their core, today’s cars consist of an extremely complex set of subsystems connected across a labyrinth of different digital connection buses, a type of interface, with names like the CAN bus, FlexRay, LIN bus and even Ethernet (yes, that Ethernet). Each of these operate at different speeds, carry different types of data, and enable connections across different parts of the car. Conceptually, it’s not altogether different from how all the different components in a smartphone or PC are connected and work together.
However, where a smartphone or PC has a handful of essential subsystems, today’s cars often have over 150 different subsystems—one for adjusting the seats, another for the steering, yet another for the brakes, and so on. Each of these subsystems is typically controlled by a device called an ECU or Electronic Control Unit, which is powered by some type of computing component, such as an ARM-based processor or an Intel CPU (or microcontroller).
Interestingly, part of the reason why cars have so many different subsystems—in addition to the fact that they’re extremely complex machines—has to do with how cars are designed. Car companies often have different teams working on different subsystems (some of which are shared across multiple car models or even brands) and each of them drives the creation of the ECUs. In fact, one former auto industry executive joked that each group felt they had to have their own ECU in order to justify their existence.
In real life, however, this is problematic. In fact, this is one of the reasons why advanced car makers like Tesla, as well as pure tech companies like Apple and Google, have such a huge opportunity to impact the automotive market. There is a sense that the market could be ripe for a simpler and, therefore, more disruptive approach to car design. Practically speaking, this is also helps explain why automotive-related technology developments often take much longer to get to market.
On top of all the hardware components, today’s cars are running an enormous amount of software. Mainstream cars may have up to 10 million lines of code and high-end luxury sedans can have nearly 100 million—that’s about 14 times more than even a Boeing 787 Dreamliner jet. Of course, given all software and that potential for bugs and security holes that it implies, it’s no surprise that a variety of companies, such as Karamba Security, are creating solutions to make cars more secure.
As impressive as the technology of today’s cars may already be, the goal of bringing more autonomous capabilities to the cars of tomorrow is driving even more innovation and more technology. Adding more types of cameras, sensors such as LIDAR (Light Detection and Ranging) and technologies to process these new types of input, such as nVidia’s artificial intelligence (AI)-enabled DrivePX, are adding even more computing horsepower to today’s cars.
Another key element is connectivity. At the recent Telematics Update Automotive trade show in Detroit, Qualcomm (a client of mine) debuted a new modular Connected Car platform that offers support not only for 4G and WiFi, but Bluetooth, DSRC (Dedicated Short Range Communications—a standard supported by many in the auto industry for vehicle-to-vehicle communications amongst cars) and even Gigabit Ethernet.
Given the interconnected nature of today’s cars, it’s also possible to do what’s called drive-by-wire. While the concept sounds scary, it’s essentially the ability to deliver car control messages to the appropriate ECUs across the various interface buses in order to help pilot the car. One of the key goals in the near term is to create safety-focused features (such as the automatic braking mentioned earlier) in new cars. However, it’s also possible to leverage these connections to create add-on kits for existing cars that can bring some level of autonomous control to them. Startups like Cruise Automation (recently purchased by GM), in fact, are working on these kinds of solutions.
Given all the computing now happening inside a car, it’s easy to see why so many companies in the tech industry have become so enamored with the automotive business. Throw in the existing but expanding opportunities for the car’s infotainment system and the automotive opportunity for tech companies becomes even more compelling.
As with many tech-related developments in the modern age, there are some additional risks that open up from the increased use of digital technologies, particularly with regards to hacking. But cars have been headed down the computer-controlled path for a very long time and the potential opportunities for improvement, particularly with regard to safety, will continue to make the increased use of technology in today’s and tomorrow’s cars, an exciting thing to look forward to.
USA TODAY columnist Bob O'Donnell is the president and chief analyst of TECHnalysis Research, a market research and consulting firm that provides strategic consulting and market research services to the technology industry and professional financial community. His clients are major technology firms including Microsoft, HP, Dell, and Qualcomm. You can follow him on Twitter @bobodtech.