Bay-Area based startup Drive.ai is exploring solutions to one particular problem that arises when self-driving cars take the human out from behind the wheel. As much development has gone into figuring out how these cars will drive safely, relatively little has been figured out in terms of how these cars will communicate with non-passengers – including pedestrians and other drivers. Once subtle, specific, human communication is removed from the equation, simple blinkers won’t even come close to covering the range of necessary communication. This includes indicating to pedestrians when it is safe to cross a street, communicating situational awareness to other drivers, and a host of other exchanges indicated using eye contact or waving. Such essential conveyances are taken for granted until the human is removed from the equation. Roboticist Carol Reiley, co-founder and president of the company, elucidates the problem: “We need to be able to communicate in all directions, and we need to be able to show intention and have a conversation with the other players on the road”.
Much like the rest of the self-driving car industry, drive.ai is using deep learning to teach its AI how to communicate with the world around it. For the most part this problem has remained untouched by the rest of the self-driving car industry. Drive.ai is testing vehicles with roof-mounted billboards designed to show messages to everyone nearby. They have explored emoji based communication to get around language and literacy obstacles, as well as a library of “safety sounds” designed to express either an aggressive posture or an intention to yield.
As good a solution as this may sound, the matter is complicated by the difficulties of attempting to find universal methods of communication that don’t involve human gestures. Signals that seem universal may not carry over for everyone, or for every culture. Furthermore, dozens of different companies are working on self-driving vehicles, all of which could easily end up with incompatible, or even contradictory, “languages”.
John Rousseau, executive director at design firm Artefact, has worked with Hyundai on a self-driving concept. He argues that a better approach would involve tying this communication into the wider infrastructure such as traffic lights and the driving environment. “Get past the immediate, this is what we can make this device do, Think about the entire system.” Rousseau says. For the moment, however, Drive.ai will work to figure out how to allow individual vehicles to communicate effectively with the world around them.