Detailed Google self-driving car

Gabe Nelson, a well-mannered autopilot of American Auto News, said after taking a ride on Google’s self-driving car, he said “it's no different from a taxi ride. It’s even possible to completely relax and choose a comfortable location to see what’s going on along the way. Landscape.” Like Google’s official tweet released at the end of April, its self-driving car has been tested on urban roads for about 700,000 miles. It can comfortably deal with urban traffic that is 100 times more than the highway, and escapes suddenly. Pedestrians and other changing traffic conditions are a great set.

Gabe Nelson also pointed out that urban travel is much more subtle than on the highway. Its rules are more complicated and traffic lights are also limited. However, it is not easy to grasp the driver’s self-determining ability to traffic conditions: the accelerator pedal is not fast, the brakes are not fierce, the body language of the cyclist when turning, and some acquiescing rules when changing lanes are often Unwritten regulations in urban transport. However, according to the current situation, Google’s self-driving cars are quite satisfactory in this respect, and are considered to be barely passable.

Although Google’s self-driving cars should nominally be controlled entirely by computers, California’s law ensures that the car is functioning properly and that emergency measures can be taken in emergencies, requiring Google to have at least one driver in its autonomous vehicle. It also created the role of "Safety Driver." Every day, they need to take a vehicle to record and track data in real time, and provide engineers with detailed feedback and reports for later improvement.

Can't open a car robot in Mountain View, this is a problem!

But it is really "love each other radish vegetables." According to Alex Davies, a journalist for Wired magazine, taking a Google self-driving car and wandering around Mountain View for a few laps (about 3 miles), the process is really impressive - a bit boring. He said:

In addition to several unskilled moves, the whole process is like the old driver driving. Even you may feel that sometimes you are too cautious, and it is no different from the young man who got a driver’s license a few weeks ago.

However, this modified Lexus RX450h is equipped with a wide range of sensors, cameras, and radar systems, but this outfit does not seem to be more advanced than the auto-driving car that preceded it.

Frankly speaking, the key advantage of Google’s self-driving car is that it does not collect data through the camera, sensors, and radar during driving, and takes corresponding measures immediately after analysis. It is more based on responses from a large amount of information on each road that Google’s automated driving team has previously collected. Before Google’s autonomous vehicles were to open up new frontiers, the team’s engineers would collect detailed features in the area including lane markings, the exact location of the curbs, the height of traffic lights, and speed limit values.

Andrew Chatham, who is responsible for map development on the Google’s automated driving team, stated that “digital maps are crucial for Google’s self-driving cars, and more detailed maps can make the design of computer control systems easier.” Because of the With the assistance of the car's central control system, it is very clear what kind of situation the road may appear, and the software can fully focus on dealing with other vehicles, pedestrians, cyclists, buildings, and emergencies.

Therefore, this seemingly "magic" map actually limits the driving range of Google's self-driving cars. As Chatham pointed out, "If we don't map out map data in a certain area in advance, then the autonomous car will be 'clumsy' there." Although as sensor technology matures, it may become less and less precise. Maps data support, but Chris Urmson, project leader of Google’s autonomous driving technology, still stressed that the role of “map” is critical.

The problem of battery capacity is the main barrier that electric vehicles cannot be widely put into the market at present, so it seems that “map” has become a “nightmare” for the commercialization of Google’s autonomous driving technology.

The problem of battery capacity is the main barrier that electric vehicles cannot be widely put into the market at present, so it seems that “map” has become a “nightmare” for the commercialization of Google’s autonomous driving technology.

This way of collecting data was wrong," although the FCC investigation report later stated that Google Street View did not collect illegal Wi-Fi data. However, this "unintentional" Google really cannot help but beware of.

Google is now vigorously developing its self-driving car technology, using the same sensors, radar and cameras. Google told us how cool this thing is. Just enter the destination in the navigation system and remove your foot from the pedal. The car will automatically send you in the past. Everyone does not think that something is wrong?

In fact, Google’s autonomous driving technology represents Google’s change in its control of user behavior. From the very beginning we control our portable mobile devices and now control the vehicles we rely on daily. Such cases are more evident in the upstarts such as Tesla. Previously, "The New York Times" reporter disclosed that Tesla uploaded all data of the user on the on-board computer system to the company's server; last year Tesla forced users to install a software update after the battery pack fire broke out. It is said that "the height of the suspension can be automatically raised when the Tesla Model S is driven on the highway."

The more exotic French car company Renault, which provides battery rental services while selling the Zoe electric vehicle, but according to German media DerSpiegel revealed that Renault can remotely control if Zoe users do not pay for the battery rental on time. The battery pack stopped working, making the car difficult to move. Therefore, as the technology of the car networking becomes more and more mature, the core functions of the car and the data of the car terminal can be mastered by car companies, governments, and even hackers.

Of course, Google can never be satisfied with the existing development model. Last year, Google Ventures spent $250 million on Uber, and Google Maps after the May 6th upgrade this year added Uber's entrance to the Uber taxi application when it provided a reference to public transportation. It is likely that Uber will receive Uber in the future, and self-driving cars will also join Uber's taxi service. However, according to Bradley Berman's point of view, Google is likely to be "hands and feet" in the algorithm of route for self-driving cars, because based on the large amount of data collected by Google and other companies, it is very likely that auto-driving cars will not choose the most shortcut or the most Safe lines, but to take the best operating mode for the company. I am afraid that after we give up control of the car, this is just the beginning of a "nightmare".

summary:

Bai Juyi's "Spring in Qiantang Lake" contains the phrase "Squandering flowers for attractive eyes", which can properly describe the current development of Google's self-driving cars. From the start of the project in 2009 to the realization of the operation of urban roads this year, it has been a long product cycle. However, the “outsiders are watching the bustle and the insiders are watching the doorway.” Google’s self-driving cars have indeed achieved very intelligent control and are more cautious in the course of driving on city roads. They can also rely on sensors, radars, and probes when encountering unexpected situations. Assisting in “changing risk into safety”, Google’s tech giant seems to go further than other auto makers.

But even so, it can't hide the "hard injuries" that exist in Google's self-driving cars.

The biggest problem first is that this vehicle is completely supported by a large amount of map data, which means that it is not currently "autopilot" in its full sense, even in areas that are completely unfamiliar, even with powerful radars, sensors, and cameras. Equipment, it is also difficult to move;

Secondly, Google’s fascination with personal user data’s “greed” may also cause future development of self-driving car technology to diverge. With constant emphasis on data security and resistance to illegal monitoring, Google, as a search engine tycoon, should pay more attention to the protection and proper use of user data.

Again, autopilot technology can easily give people an emotional sense of substitution. How to enable users to rest assured that the steering wheel to the computer control, as a provider of terminal technology, Google how to ensure that users can enjoy the convenience of science and technology at the same time, not because of the third party and cause damage to their own security and interests.