How to make driverless cars to understand the words of pedestrians?
-Driverless cars, Mercedes-Benz-IT news How to make driverless cars to understand the words of pedestrians?

Last month unveiled Google driverless car let us shake the steering wheel in the automotive space, it allows us to free their hands, do other work, entertainment activity. But this does not mean that vehicle around pedestrians and moving is no longer needed in a driverless car "communication" – ever, ever stop on pedestrians and other links such as, the driver can make their own judgments and direct communication with the pedestrian, but driverless cars how to solve this problem? Or, when we need to take the car, often free to point your fingers at the roadside location, so car to pull over. Future driverless taxi how to receive the order?

A few days ago, Mercedes-Benz and robotics experts and linguists consider ways to enable Exchange of driverless cars and pedestrians, it offered a solution-using hand gestures.

Mercedes-Benz used the action tracking system. It and electronic arts laboratory for the future cooperation in the four-axle vehicles move on a research action tracking system, and other auxiliary systems can translate gestures of pedestrians into the car with the language. For example, in one experiment, Mercedes-Benz made a four-axis aircraft, "understand" has taken off, stop, rotate gestures, and Mercedes-Benz says, this technology may be embedded in portable devices in the future, so that you can achieve "understand" human sign language.

In fact, some simple technology to communicate with cars have been able to achieve. Through networking, driverless cars could communicate, share information, navigation routes, traffic, accidents, congestion, and even informed the cheapest nearby gas stations. Mercedes-Benz today's research, still faces hurdles.

First hand learning costs too much--whether it be the car people to learn gestures or people studying human gestures, car understands sign language, the biggest obstacle. Language more straightforward than gestures, so that ordinary people will not care too much that is mainly used in domain-specific gestures. Also, localized gestures were too big and would increase the costs of learning. For example, some places use the thumbs up hitchhiking, while some applaud this gesture, palms forward and gently side to side instead of greeting and get a ride.

Secondly, in the congested crowd, driverless cars how to identify who it was and it "speaks". The same gesture may have errors, not to mention different signs.

Is still the most important and biggest concerns of security issues. Cars know that human action using gestures on behalf of meaning, but people also need a car how strong safety awareness. We have to understand clearly the time to walk on the sidewalk, next to driverless cars would not because "poor security-conscious" and hit. This car on the reverse communication has still not been solved.