Google is taking autonomous technology to a whole new level, as the company's latest self-driving car monthly report says that the project's team has been teaching its cars' AI prototype how and when to honk on its own.
The report, which was released on Wednesday, notes that engineers involved in the project have been programming the cars' honking algorithm to know the most appropriate time to honk as well as use different honks and beeps for different situations.
Such circumstances include a car backing out of a blind driveway and a car going down the wrong way on a one-way street. In these scenarios, the car would first give a small honk inside the vehicle so that the engineers could record whether it actually needed to honk and then provide feedback to the AI. They will then let the AI honk on its own once they believe it is ready.
"Our goal is to teach our cars to honk like a patient, seasoned driver," Google's self-driving car team wrote in the report. "As we become more experienced honkers, we hope our cars will also be able to predict how other drivers respond to a beep in different situations."
While the autonomous car will belt two short pips to another vehicle that is slowly reversing toward it, the vehicle will also let out one loud honk in more serious situations.
Google has also given its prototype cars a "hum" so that they could mimic the sound characteristics of traditional cars, giving pedestrians and cyclists around them the chance to hear them coming. The hum will also increase when the car speeds up and decrease when it slows down.
The engineering team has experimented with a variety of tones to create the basic sound, such as those of ambient art sculptures, other vehicles and consumer electronics products and even Orca noises. However, the process of finding the perfect sound continues.
