The map app can predict traffic lights, and it's really not a black technology

I believe that many poor friends who drive have recently noticed a detail, that is, when waiting for the traffic light, the navigation software will give you a countdown to the red light reading seconds at some intersections.

To be honest, at first, Shichao felt that this function was a bit chicken.

After all, the signal lights at most intersections do not themselves have their own seconds reading.

But after a while of experience, I found that it was quite practical.

For example, after passing the first intersection, the navigation software will immediately release the advance warning of the signal light of the next intersection.

Without this tip, I would most likely want to kick the throttle, but in the end it would definitely be fuel consumption and wallet that hurt.

Of course, what the editorial office can't put down the most is the voice prompt 5 seconds before the green light comes on. This is definitely a boon for everyone who loves to swipe your phone when waiting for a red light (it is not recommended to drive and play with your mobile phone).

But every time I see that flashing number in the navigation interface, I think about the question, how exactly does the software's second count work?

When Shichao collected information on the Internet, it found that the answers to this question were also divergent.

Some people said that what is the difficulty of this, the traffic management department will take action.

There is also a certain basis behind this statement.

For example, in January this year, the Xi'an traffic police connected data to Baidu Map. In addition to the status of your lights, you can also receive information on traffic accidents, road construction, traffic control, and more in the app.

It's like putting a God's perspective on the driver, you just drive, and the navigation does the rest.

However, this seemingly perfect solution is really difficult to implement on a large scale.

First, the traffic management systems of different cities are all separate, and trying to rub them all into a platform is obviously a difficult system project.

Second, even if you are connected to the traffic management system and want to obtain the second reading data of the traffic lights, you have to pull all the signal lights together to "surf the Internet".

Or take Xi'an as an example, they replaced nearly 2,000 traffic lights in the city with smart signal lights, and then there was this wave of operations.

Considering that there are still many traffic lights across the country that are still "offline", plus temporary signal lights will be used in some places during the morning and evening rush hours. The navigation software simply wants to rely on the data of the traffic management department, which is somewhat inadequate.

Then, if you want users all over the country to experience this function, software manufacturers also have a housekeeping skill, that is, algorithms and big data.

The essence of second prediction is actually not complicated, in fact, it is a process of analyzing data and finding rules.

The principle of the signal light is to make a cycle of changes according to a certain rule set in advance.

Then, what the software needs to do is to calculate the change period of these signal lights with the help of the data provided by the user.

In the national invention patent applied for by this Amap, a solution is disclosed.

First, they retrieve the information of the phone's background acceleration sensor and refer to the actual current location of the vehicle.

Then by judging whether the vehicle went from stationary to starting, you can infer whether you are waiting for a traffic light at an intersection.

With this conclusion, the navigation software can count the number of traffic lights passed by each vehicle and the time interval at which the vehicle starts in a time period.

Put this data into the same coordinate system, and you will find that they are like sinusoids with a regular change.

Finally, through simple algorithmic processing, the time change of the signal light can be seen at a glance.

However, because the algorithm "prediction" relies on user data to eat, it also leads to certain limitations in the current second reading function.

For example, some intersections do not have many people passing by, and navigation software can obtain less data. Once the software company lacks food, it can't help...

Therefore, in the actual use process, you will find that not every intersection will appear to read the seconds.

In addition, compared with the solid data of the traffic management department, the open reading seconds obtained by the algorithm will produce some errors.

Shichao often encounters the situation that the software reads the seconds several seconds faster than the actual one. If you trust navigation too much, you may be happy to get a ticket.

Although the current software is not perfect, its potential is huge in the long run.

Everyone must have encountered this situation when driving, obviously they have been racing all the way, but every time they come to the intersection, they are always "just" blocked by the red light.

It's certainly not that the signal lights are meant to embarrass you, but because you haven't found a proper driving rhythm.

At this point, the algorithm will tell you everything.

Since the navigation software has real-time data on each traffic light in your route, it can automatically calculate what speed you should maintain for each road section, thereby improving the pass rate of intersections.

I believe that friends who have walked through the green wave section should understand how refreshing it feels to be released by a green light at every intersection.

But the efforts of software companies may be the next bigger move.

The "vehicle-road interconnection" we mentioned in past articles relies on the deep combination of software and hardware.

For example, navigation software can share the results of real-time estimates with the traffic police system, let them know which road is currently facing severe congestion, and tell traffic lights how to adjust the law to improve the efficiency of traffic, thereby saving everyone's time.

At the same time, it can also work with the autonomous driving system, the vehicle provides traffic data, and the software makes decisions for the vehicle, letting it know how to choose the route and how fast it should use.

In other words, it can achieve a certain degree of "smart transportation" in a relatively low-cost way.

It is foreseeable that when new technologies become more perfect, our way of traveling will become safer and more efficient.