“Having some collaboration from the infrastructure is very, very helpful.”
The great lie that technologists tell the public is that magic happens overnight.
Magic doesn’t exist and progress rarely happens overnight.
Let’s take the driverless car as an example. Sometime in the next decade (or, last week, if you want to believe Tesla founder Elon Musk), a tech executive will stand on a stage and present the first “driverless” car meant for mass public consumption. The spin will be that this new innovation was created solely by the executive’s company and that it’s magic was born of some great inspiration by some entrepreneurial genius who will then grace the covers of Fortune and The New York Times.
The fact is: technological progress takes a lot of time and pragmatism. One block is built and expands into the ecosystem and then another block is built on top of that.
The future of the self-driving car will work exactly in this way. We have already seen how the Mobile Revolution has provided the basis for the technology for the driverless future. But building a fully autonomous ecosystem is not that easy. The car’s hardware, connectivity and the infrastructure of the world all need to work hand-in-hand to create the dream of a driverless utopia.
Companies, organizations and governments from all over the world will need to contribute. The Mobility Transformation Center at the University of Michigan has 60 partners that include chipset manufacturers, infrastructure providers, insurance companies, law firms. Even Xerox, which is the largest maker of toll collection equipment in North America.
“When cars leave their intended path, they don’t want them to hit objects. But in the real world they will hit objects. They need to avoid those objects and safely navigate around them,” said Carrie Morton, deputy director of Michigan’s MTC. “To really come together, what the university recognized early on was that not just one sector was going to be critical in moving this transformation forward. We couldn’t do it, for example, without having the insurance industry on board and understanding how these new vehicles will be insured, what will those new models be?”
Intelligent Hardware And The Connected Car
Tesla provides an interesting case study into the future of the autonomous and driverless car.
At this point, Tesla seems all in on the concept that a car can and will drive itself without any assistance from the world around it or a human driver. Tesla announced this week that all of the cars coming off of its manufacturing line will be equipped with thousands of dollars of sophisticated sensors that will, eventually, let it be a fully autonomous, self-driving car. In this case, the car does all the work like reading road signs and lines, understanding obstacles like humans or debris and sensing its environment in different kinds of weather.
Tesla is trusting the sensors, hardware, on-board computers, neural network and algorithms to do all the work for the car to drive itself. If we take that all of those components as a single vehicle unit, sure, a car can drive itself. Google has already proven that it will work, in limited scenarios.
The next aspect to consider is the connectivity of the car. Basically, think of a car like a much larger smartphone that can move around. Cars will be equipped with Wi-Fi and cellular technology like LTE. With GPS and Bluetooth. With LiDAR and radar. The automotive industry even has its own band of spectrum (from 5.850 GHz to 5.925 GHz) that will be, theoretically, only used for vehicle-to-vehicle and vehicle-to-infrastructure communications.
“We use spectrum for lots of different purposes,” said Dr. Huei Peng, director of UM MTC, in a briefing with reporters, including ARC, at Michigan’s campus in Ann Arbor. “Some of them are treated as untouchable such as military or government use. Then there are things that we think are very important but are now on the table, in review, which is fair. The only question is whether using the same channel for both Wi-Fi and vehicle safety application is a good mix. Or will there be any interference? In which a few dropped packets will be the difference between life and death.”
Autonomous Cars Will Be Better With Connected Infrastructure
The connectivity of the car is important. The pragmatism of building blocks that will create the autonomous ecosystem relies on the gradual deployment of cars that can communicate with each other. Think of automotive mesh network that let’s every car know where other cars are and what they are doing. At this point, that mesh network does not exist, which makes it much more difficult for autonomous cars to drive without the intervention of a human.
“We constantly talk about the Google car or the Tesla car can see everything that they need to see,” said Dr. Peng. “The issue though is what around the corner in an urban condition? The line-of-sight condition is generally not available. But communication will be able to let you see what is around the corner. You can see that little kid behind the bus, but he is going to run out from behind the bus. All the onboard sensors will have difficulty understanding the situation. But communication can make it much more reliable, much more cheaper, to understand the situation.”
Herein lays the rub. The connected car becomes better with more connected cars on the road. As of this point, there are very few vehicles on the road that are connected in such a way. Even fewer if you take into account cars that are connected and have a degree of automated features. Connecting cars is not the hard part (just plug a dongle with wireless properties in to a car’s OBD II port). Creating the ecosystem of cars that work in concert is what will take time.
“We also don’t think that the autonomous vehicle should try to have all the intelligence on board,” said. Dr. Peng. “Having some collaboration from the infrastructure is very, very helpful.”