April 6, 2017
I talk a lot about autonomous or self-driving vehicles. I believe it is one of the top paradigm shifts this society will see, and it is fraught with a plethora of wide and deep subjects. The eventual influence of unattended driving will make a significant shift in many of the transportation archetypes we currently have. Can you imagine that someday, all vehicles will just be nodes on a transportation infrastructure, with any and all types of activities going on inside of them.
I can only imagine what society will be like when people see today’s vehicles the way today’s teens see wired telephones.
But, that is many years off, and I digress. Back in the real world, presently the picture of self-driving vehicles is comprised mainly of driver-assist technologies. While hands-off driving is a possibility today, it is still only in beta test, regardless of what Google and Tesla claim.
However, coming down the pike is a set of technologies that will ratchet up the hands-off scenario a couple of notches. These will be the core technologies that, once all the elements are in place, will enable true driverless vehicle infrastructures.
They are a class of communication technologies that are two-way. Under the generic term V2X, they include vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), vehicle-to-grid (V2G), vehicle-to-home (V2H) vehicle-to-pedestrian (V2P), and others that are yet to be find their acronyms.
As these evolve, they will replace what is currently the state of the art – sensors. Most of what is going on, right now, in this space is equipping vehicles with various types of sensors and sensing equipment. This includes ultrasonic sensors, cameras, and various renditions of radar for the most part. And equipping pedestrians, infrastructures, the home and the grid are the more complex of the set. The one getting most of the attention now is V2V.
Why? Because V2V is going to be the easiest of the set to implement. Car manufacturers are on board, the Department of Transportation is on board and dedicated short-range communications (DSRC) players are on board. So this is one of those rare occasions where all the players are on the same team.
DSCR is really an ideal technology for smart cars. With DSRC, V2V communications can contain all sorts of pertinent information: location, speed, direction, braking status and more. That data would be updated and broadcast up to 10 times per second to nearby vehicles. That is whole lot better than relying solely on sensor data received by each vehicle. Such data allows the vehicle to make not only sensory decisions but informational, as well. Couple that with advanced artificial intelligence and this is the biggest single current improvement in fully driverless cars because it involves the ability to receive data from other moving targets – orders of magnitude more valuable than just bouncing signals off them.
Information from the rest of the V2X vectors will, of course, be useful, but most of it will come from static environments where change isn’t nearly as significant as targets moving at speeds of up to 85 mph today, and who knows how fast tomorrow. Plus, the V2V environment is constantly changing, in real time, inside extremely short time frames.
DSRC can “see” about 300 meters in any direction. There is still a bit of fear, uncertainty and doubt (FUD) around V2V, one of which is the computing horsepower required to assess, analyze and act on myriad sensor data and DSRC communications coming in simultaneously. Such systems will need to handle 1,500 or more request per second, and that is still a challenge for system hardware, today. Once the full V2X ecosystem is in place, that number could easily double. And it isn’t just the inputs. At high speed, changes occur in fraction of a second so the hardware must be capable of extremely fast analysis and computing time. Finally, add redundancy, security, and error checking to all of this and it poses a real technical challenge for smart vehicle system designers.
Not that it won’t happen, but it will be awhile. Still V2V is a strong jumping off point. And the other V2X technologies are on parallel tracks, so it’s not like they will have to be linked together in the development chain. But once it all comes together, it will be a game changer.