Your Car Will Soon Be Smarter Than You

We all know that self-driving cars and other autonomous vehicles are coming. Prototypes of self-driving vehicles can be seen around Silicon Valley, and self-driving features are commercially available in Teslas and other brands. However, there are significant “smart car” features on the way that will affect both human- and self-driven vehicles.

The current generation of self-driving vehicles is implemented using machine analogs of the human senses. A self-driving car today “sees” the cars around it and “sees” the road signs, road markings, and pedestrians that might be entering the drive path. But the next generation of cars will leverage information made available through a mesh network created by the cars themselves.

The trend toward self-driving cars has significant regulatory hurdles. But there are also regulatory incentives driving (no pun intended) the move toward smarter vehicles.

One of the key initiatives is a new standard that would require the implementation of vehicle-to-vehicle (V2V) systems in all new vehicles.

A V2V system broadcasts and receives information about speed, direction, indicator status, brakes, and other information.  We humans use brake and indicator lights to communicate with other human drivers. The V2V system will have clear advantages over signal lights: It can offer additional information (I have a flat tire) and can be “seen” around corners. V2V can also use “mesh” capabilities to advise of traffic problems in the distance—traffic slowdowns will travel from car to car across a busy highway to give you real time notification of problems.

Vehicle-to-infrastructure (V2I) is a related technology standard that would allow vehicles to receive and transmit information to a central “infrastructure.” This could include road construction advisories, availability of parking, real-time weather, and so on.

Even further, some are talking about V2X (vehicle-to-everything). V2X includes V2P (vehicle-to-pedestrian) networks that would allow cars to be aware of pedestrian movements. Your smartphone would broadcast your movements. As you walked down the street, cars in that street would slow down to take into account the possibility that you might walk onto the road.

Hacking an autonomous vehicle is an obvious concern. V2V systems are required to be firewalled so that you can’t use them to control a remote vehicle. However, the potential for abuse is obvious: If I can broadcast fake V2V signals, I could force a car to take avoidance action. For instance, a carjacker could broadcast a V2V signal indicating a pile-up, forcing a car to stop.  Spoofed V2V messages might be hard to screen—since reaction times for autonomous actions such as auto-breaking have to occur within milliseconds.

V2V, V2I, and V2X are implemented by dedicated short-range communication (DSRC) devices. If the proposed standards come into force, DSRC devices will become mandatory in new cars within the next 5 years. The result will be a massive new entrant into the Internet of Things—millions of vehicles broadcasting their status in real time.

V2V and V2I technologies will make driving safer, even without the introduction of autonomous “self-driving” vehicles.  However, as these technologies emerge, they will reduce the technical obstacles for self-driving vehicles.

Autonomous vehicles are expected to be massively safer, but it’s not all good news. There are at least 5 million jobs involving driving in the U.S. alone, and many ancillary jobs (service staff at truck stops, etc.) that could be lost. We also need to work out the legal liabilities and ethical conundrums posed by autonomous vehicles (how do we program a car to choose between two “bad” outcomes?).

However, as with most technological revolutions, this one seems unstoppable, so we had best confront the issues rather than hide our heads in the sand. For data scientists, V2V will open up a new world of traffic optimization challenges. For programmers, the emergence of autonomous vehicles highlights the need for autonomous programs and AI to develop an ethical sensibility.  And, for anyone who drives for a living, now is a good time to consider a career change.