For the gift, start research and development of graphene batteries

Chapter 96 The report is not easy
Jiang Ming got the task book and returned to the office happily.

The next job is very clear, which is to reconstruct the software architecture of the autonomous driving system.Make the software architecture as perfect as possible, including more interfaces.

Regarding this, Jiang Ming felt a little stressed.

Although the knowledge of automatic driving technology in the technology tree system is inherited, it still takes a lot of brains to apply it to specific projects.

In fact, the quality of an automatic driving system depends more on the quality of the software architecture.Get it right in the general direction, and the rest will be much simpler.

Jiang Ming brought back the project's requirements document and conducted research.

The first is the functional interface of the software, including hundreds of interfaces.There are low-level device interfaces, function definitions including cameras, etc., as well as internal data transfer and allocation functions, and there are many kinds of them.

First, Jiang Ming listed all the functional interfaces, and then listed it as an item to be developed.

The next step is to sort out the framework of the entire software and clarify the logic of program execution.

This item is actually very important.

Compared with the pure computer version of the operating system, the autonomous driving system is more appropriately called an embedded system.

Embedded systems focus on the combination of software and hardware to make better use of hardware resources.

Therefore, when sorting out the logic, Jiang Ming needs to sort out the sequential data of the equipment running in the software to prevent conflicts between the equipment.

At the beginning, it must be the self-test of the system to ensure that the data status of the system itself is normal.

Including clock, memory, interface devices, etc., a very complicated self-test process.If you have a problem yourself, an error will be displayed in the corresponding data flow.

If there is no problem in the self-test, the system can run normally.This point is similar to the process of common desktop operating systems.

Next, there is the transmission and analysis of data streams, which is similar to ordinary embedded software.It's just that the frequency is much higher than the general software.

In other words, it is the big step of modeling, which is the core of the entire automatic driving system.

The system needs to identify the signals transmitted by the camera, then draw them into a scene map, distinguish the collision objects, normal markings and other meanings, and then calibrate them.

This step is very difficult. The system needs to know what kind of object means what it means.Otherwise, the label is wrong, and the system cannot make a correct judgment.

In this step, it is necessary to connect to the Internet, call the data of the central processing unit, and perform final calibration according to the feedback of the central processing unit.

Although the vehicle has twelve camera images, for image processing, there is an underlying processor to assist in processing image data, so the amount of data is not as large as imagined.

Although not large, there are many types. The main program needs to mobilize the functions of subroutines to process the transmitted data and give preliminary suggestions.

Finally, after a series of complex judgments, a modeling diagram is generated.

The data that needs to be judged and processed is very complicated. Fortunately, Jiang Ming does not need to sort out the process one by one. Jiang Ming only needs to clarify the context.

Jiang Ming understood the logic, and then drew the main line.

Another point is about the construction of micro servers.

Jiang Ming needs to define the type of data that needs to be transmitted between vehicles, and determine what kind of judgment is made based on which data.

In addition to the main items, Jiang Ming listed hundreds of wind items.

The logic diagram at this time, densely packed with fulcrums, looks dizzy.

Even Jiang Ming, who has mastered the inheritance of autonomous driving technology, took a long time to sort out the logic.

Thanks to Jiang Ming's Peiyuan Pill, this can ensure that the body's functions are sufficient, and he will not faint due to overwork.

Even so, it took Jiang Ming nearly a week to complete.

It was two days longer than the five days Jiang Ming had expected.

Because this architecture is for software engineers, and it will be reported to Qin Ruya.Therefore, Jiang Ming needs to make a complete report.

It includes the entire architecture diagram, network MAP diagram, control logic, and the definition of network nodes for all interfaces.

Especially regarding the deployment of micro-servers, Jiang Ming spent a lot of time.

Although he was very tired, after printing out the information and reading the thick book, Jiang Ming felt quite fulfilled.

On the eighth day, Jiang Ming took the report and prepared to report to Qin Ruya.

"Jiang Ming, what have you been up to lately? Why haven't you seen your people?"

When going to Qin Ruya's office, a hardware engineer saw Jiang Ming and asked strangely.

"Busy on software architecture issues."

Jiang Ming replied.

"Software architecture problem? What's wrong with the software architecture, isn't it used well?"

The visitor asked strangely.

"Uh? Isn't the software for autonomous driving going to be re-architected to suit the model?"

Jiang Ming is also very strange, a little bit wrong.

"Software re-architecture, who did you listen to, we have not been notified. Part of the software for the autopilot system has come out, and we are currently testing the sub-items."

said the engineer.

"Huh? Part of the application software has already been released. It's impossible. The overall structure of the software has not yet been released. How can there be an application program on the Internet?"

Jiang Ming was puzzled.

"Here, look, this is our latest mission statement. I am in charge of debugging the distance sensor, and the application code has been given."

While speaking, the younger brother handed Jiang Ming his task book.The above does show that the application code needs to be tested.

"Isn't it, so fast? So colleagues in the hardware group have been testing applications recently?"

Jiang Ming asked.

"Yeah, I have a lot of work recently, and I am a bit busy. How about you take the test with me, so that both parties will be more relaxed."

The little brother suggested.

"Wait for a chance, I have something to ask Director Qin."

Jiang Ming refused.

"Okay, come and help when you have time."

Afterwards, Jiang Ming trotted all the way to Qin Ruya's office.

"Jiang Ming, you're here. It's just in time. There's a plan for hardware testing. You can go do it."

Seeing that Qin Ruya was very busy, she looked up and saw that it was Jiang Ming, then continued to look down at the documents in her hand.While watching, he said to Jiang Ming.

Seeing that Qin Ruya didn't care about the software architecture, but instead arranged for hardware testing, Jiang Ming was a little speechless.

How forgetful Qin Ruya is, she forgot the work that was arranged just last week.

Or is it that Qin Ruya doesn't care about the work of software architecture at all?
"Sister Qin, it's like this. I heard outside just now that the automatic driving challenge project has started the application test. Isn't it a bit fast?"

Jiang Ming was not interested in hardware testing, so he said so.

"Is it fast? No, time is running out. We need to speed up the progress."

Qin Ruya said.

"However, the software framework of our driving system has not been built yet, so how come we have reached the stage of application testing. In this way, the adaptability of the system will be greatly reduced. And there will be problems in the subsequent program transplantation. "

Jiang Ming was in a hurry and said this.

"Software architecture? Hasn't the software architecture been determined, is there no problem?"

"No way, Sister Qin, you just arranged for me to do the software architecture last week and gave me ten days. It's only the eighth day. You tell me that the software architecture has been determined?"

Jiang Ming was a little impatient, "Hey guys, I worked overtime to make the software architecture, and you used someone else's before you even looked at it. This is a bit inappropriate."

Jiang Ming thought that someone cut Hu halfway, so he said so.

At the same time, Jiang Ming was secretly startled, whoever made the move was much faster than himself.

Jiang Ming did not expect that Qin Ruya and the others did not make a new structure at all, and completely followed the previous one.

"Forehead?"

Hearing Jiang Ming's words, Qin Ruya also reacted, as if Jiang Ming had been assigned work to refactor the software.

But I just said this because I couldn't bear to hurt Jiang Ming's confidence. Could it be that there is some new progress now.

"Oh, this matter, I remember, what, there is a result?"

Well, the time of eight days is a bit early, and the result should be almost there.

Qin Ruya thought that Jiang Ming's software architecture had reached this point and he couldn't go on, so she turned to herself for guidance.

Jiang Ming, it's not that I don't want to give you a chance, it's that the work of software refactoring is too big, and we can't handle it.

"Yes, Sister Qin, it's finished."

Jiang Ming said excitedly.

"It's done, it's done, let me see."

As Qin Ruya spoke, she took the materials in Jiang Ming's hands and prepared to make critical comments.

Not to mention, there are quite a lot of pages, which is quite heavy.

Touching the thick document in her hand, Qin Ruya thought of this in her heart.

However, no matter how thick the information is, it is useless without practical things.

To be honest, Qin Ruya was very unfavorable about Jiang Ming's work.Jiang Ming wouldn't be filling the content of the report with some detailed information, and then trying to get away with it.

However, this is for commercial purposes, not for schools. Every piece of information is useful and must produce actual value.

"Tell me your plan first."

Qin Ruya didn't open the information, but asked Jiang Ming like this.

"Okay, let me briefly talk about my refactoring plan."

"The software refactoring this time is aimed at simplicity, efficiency, and portability. Specifically: the entire interface is required to be concise, the logic is clear, the actions are executed efficiently, and it can be freely ported in different devices."

"In order to achieve these three goals, I have adopted the following plan."

"First of all, the Yin Mington logic framework is adopted. The advantage of this framework is that the interface is simple enough, and the interfaces of each device and logic are clear enough. The mutual references are minimized as much as possible, and the global variables are as small as possible. few."

Jiang Ming explained.

"Wait a minute, what's the situation with that Yin Mington logical framework you mentioned?"

Hearing this, Qin Ruya asked.

"Yin Mington's logical framework is the software logic framework first proposed by Deng~Guoyou scholars. Compared with the traditional tree structure, Yin Mington innovatively uses the fishbone method to distribute software functions. Later By means of progressive layers, logical decentralization is achieved."

The question was simple, and Jiang Ming came here casually.

Hearing Jiang Ming's explanation, Qin Ruya was a little stunned. Why hadn't I heard of this Yimington logical framework?

Although I am not a software professional, but as a person in the autonomous driving industry, I don't know all the technologies in this industry, but I am very clear.

This Yin Mington frame has no impression at all.

Qin Ruya really wanted to say, I've never heard of you, Yin Mington, did you make it up.But seeing Jiang Ming's serious look, Qin Ruya didn't open his mouth.

"Go on."

Qin Ruya wrote the word down in her notebook, and then signaled Jiang Ming to continue.

"In addition, considering the portability, I refer to the software framework structure of the industrial computer, and adopt the design concept of von Nosky cylinder modular program. When the program is transplanted, only 1% of the underlying code needs to be changed, that is, The suitability of the software can be guaranteed.”

Jiang Ming continued.

"What, 1%."

Good guy, Qin Ruya has never heard of that Von Noski cylinder modular program.But Qin Ruya is very clear about the 1% code change.

You must know that there are very complicated control systems on the market, and each manufacturer uses different ones.

In order to adapt to these control systems, the application software must be modified. This part of the problem is not a big problem, and it is done by the application engineer.

But a lot of hardware requires you to change the underlying software, which is a bit difficult.

The underlying software involves the architecture of the entire system and is related to the stability of the entire software system. It is said that one move affects the whole body and cannot be changed at will.

But often if you don't change the bottom layer, you won't be able to adapt to the other party's system.

Therefore, in the software architecture, subsequent portability must be taken into consideration.

On the one hand, it is very good to be able to control the amount of underlying software changes within 5%.

Jiang Ming's design concept can be controlled at 1%, which is a bit shocking.

Could it be bragging?
Qin Ruya thought so in her heart.

Moreover, there is such a powerful design concept, it makes no sense that I have never heard of it before.But Qin Ruya thought about it for a long time, and she had never heard of this so-called von Noski cylinder modular program.

Is it because I am ignorant?

No, I have attended so many industry conferences and listened to so many reports over the years, but I have never heard of them.

At this moment, Qin Ruya had doubts about herself.

"Zhang Ru, come to the office."

Thinking of this, Qin Ruya couldn't sit still.

Qin Ruya signaled Jiang Ming to listen first, and then called Zhang Ru to let Zhang Ru pass by.

Zhang Ru is an expert in the software of Xiake Academy. I believe Zhang Ru should have heard of such a powerful theory.

While waiting for Zhang Ru to come to the office, Qin Ruya picked up the thick report given by Jiang Ming and read it.

not simple!
Qin Ruya felt that this report was very difficult!
(End of this chapter)