Google's automatic driving car can already run at a practical level even in urban areas with complicated traffic conditions

Automatic driving car under development by GoogleAs the name implies, the future rides are vehicles of the future that the car can recognize the circumstances and run automatically without driving by the driver. It is an automatic driving car that has been repeatedly tested, mainly on expressways with relatively stable traffic conditions so far, but it has already reached a practical level by refining the demonstration driving in urban areas repeatedly It became clear from the latest movie.

Official Google Blog: The latest chapter for the self-driving car: mastering city street driving
http://googleblog.blogspot.jp/2014/04/the-latest-chapter-for-self-driving-car.html

The movie that the Google automatic driving car reaches the practical level at a glance is as follows.

Google Self-Driving Car on City Streets - YouTube


VICERA of Google Automated Driving Car Project member who will introduce the latest video of Google's automatic driving car.


I get on an automatic driving car. Incidentally,Lexus RX 450hIs a base vehicle.


This time, the automatic driving car is going to run in the city area of ​​Mountain View in California where Google's head office is.


Vicera in the driver's seat does not touch the steering wheel, but the automatic driving car clears the left turn brilliantly.


A purple box-shaped object on the screen represents a car that the automatic driving car recognizes. You can see that the signs under construction are also correctly recognized.


As a sign, I found a construction site ahead.


I will pass between pylons without difficulty.


Discover obstacles ahead, check the right and left with sensors to see if there is any problem in progress. A green ladder-like mark indicates that the course is clear to this position.


It was a pickup truck parked on the street. The automatic driving car keeps moderate distance and avoids it.


I came to the railroad crossing point. There is still a car in front of the railroad crossing, so a red ladder-like sign indicating that you can not go ahead is out in front of your car.


Wait in this state and confirm that the previous car has crossed the railway and came out of the railroad crossing ... ....


It can be confirmed that the automatic driving car has entered the railroad crossing.


The red object seen in the front right direction is the bicycle while driving. For an automated driving car, a bicycle is the one that pays the most attention among moving objects.


I inform the cyclist that I will change my lane by raising my left hand.


The automatic driving car can predict the danger by recognizing the gesture of the cycling person.


If cyclists think that they have returned to the bicycle exclusive lane ......


Also, suddenly you may enter the car lane. Automatic driving cars avoid collisions by paying utmost care to the capricious behavior of bicycles.



I came to an intersection with a pedestrian crossing. Traffic accident is the most likely situation.


A screen-like yellow object represents a pedestrian. Automatic driving car that waits for pedestrians to cross the pedestrian crossing. At this time, "Since this person has advanced a crosswalk already more than halfway, it passed OK behind that person" and "This person has just started crossing the pedestrian crossing, so I waited" and checked the current status against multiple goals I will.


All the pedestrians have gone and cleared the right front, but since we detected that the bicycle is approaching from the right backwards, we will not enter the intersection yet.


After the bicycle also passed and it became safe to pass, the automatic driving car entered the intersection.


In this way, it seems that you already have acquired the ability to avoid involvement accidents.


The Google Autonomous Car has already tested about 700,000 miles (about 1.1 million kilometers) of automatic driving. According to Google, even though traffic conditions of urban areas that seems to be complex and chaotic at first sight, assuming thousands of situations and letting computers learn, it is said that what will happen next is sufficiently predictable Thing. Google is optimistic about the realization of the future that the driver will automatically drive to the destination without any driving operation.