ALL NOVEL Chapter 568
Chapter 574
Before Lu Yu went to the physical examination, he had specially explained it.
The problem involving electromagnetic interference must be solved until he returns, and no one is allowed to deal with this scientific and technological point privately.
And Lu Yu’s prestige in the Longguo 2 aerospace center naturally does not need to be said.
He said no.
Then even if other researchers want to touch this problem with their own hands, they can’t really blatantly start dealing with the trouble in advance.
At most, I think privately about what is more appropriate to deal with.
It’s like making a plan in advance.
Lu Yu also guessed this, the researchers must not be idle, so as soon as they came to the conference room, let everyone first talk about the ideas in their minds, and then he will analyze which ideas are more feasible, and which ideas are more serious and can be ignored directly.
“Chief Engineer Lu, let me talk about my thoughts first.”
A big wave female researcher stood up and said her thoughts very briefly and competently: “If I look at this problem from my point of view, I would prefer to use the Kalman filter algorithm.”
“Everyone should be familiar with this algorithm.”
“It is an algorithm that can estimate the state of the system based on observational data and model predictions, and the specific role is to effectively deal with the uncertainty and noise present in the system.”
“And the problem we are currently facing, the Capol filter algorithm is a good solution, which can greatly alleviate the compatibility problems caused by our electromagnetic interference!”
“Think about it.”
“Why do compatibility issues happen to our microsatellite network at the end of the day? Isn’t it because our microsatellite network system is not stable enough?
“If we can make it much more stable, even if it does face electromagnetic interference.”
“At least in terms of compatibility, I don’t think there will be any problems that can comfort our entire satellite network, of course, it is not realistic to have no problems at all.”
“But isn’t the purpose of our research to avoid the problem of interfering with the launch of Zhurong-2 to the greatest extent?”
“Once faced with electronic interference, we can immediately use the Kalman filter to filter and calibrate the MEMS system output data to improve the accuracy and stability of the data.”
“You can refer to this paper, which describes how to use a Kalman filter to compensate for the error of a MEMS inertial measurement unit.”
It turns out that the female researcher put her search results during this time on the projection screen, and the content of the paper is simple, describing in detail how to use the error compensation problem of the Kalman filter, and there is no twist and turn.
That was decades ago.
During the Cold War period between the Eagle Country and the Northern Bear Country, a scientist made up a paper.
I have to say that although the world situation during that time was not stable at all, it was like a powder keg, and there was a risk of explosion at any time!
But during that time, the speed of human scientific research was indeed very fast, and even until now, many research results were dug up from the research ideas of that year.
Scientific research is not like scientific research, but like archaeology.
It can only be said that war and existential threats are the best driving force for human development.
“Well, good.”
“The Kalman filter is indeed a suitable solution.”
“Does anyone else have any other ideas?”
After Lu Yu roughly scanned the paper, his eyes suddenly showed a satisfied light, and he nodded lightly and said.
Although he did not directly read the paper, many of its ideas coincided with him, and the use of Kalman filters to solve the possible troubles of micromotor systems was also a solution that he had in mind in advance before coming.
But… Although Lu Yu praised the female researcher, it did not mean that Lu Yu had planned to use this scheme to directly solve the problem!
The Kalman filter algorithm is used to solve this problem, and there are indeed certain advantages.
But… The problem is just as obvious and cannot be ignored.
It’s just that he didn’t say it directly now.
“I prefer to use a support vector machine over the Kalman filter algorithm.”
Sure enough, another researcher stood up directly and said:
“The Kalman filter algorithm does have many advantages, but the vector calculation method has a clear advantage!”
“It can build algorithms for classification or regression models based on training data, simply and efficiently processing high-dimensional and non-linear data!”
“We can use support vector machines to perform categorical regression on the output data of MEMS systems and quickly increase sensitivity. And the vector machine to compensate the temperature of the MEMS accelerometer… Nor is it comparable to the Kalman filter algorithm!
A round-bodied researcher offered his opinion.
As soon as he said this, several other researchers present also nodded in satisfaction.
This was not just the result of his discussion alone, but the conclusion of their group after some discussion.
The advantages of the Kalman filter algorithm are indeed many.
But if vector computing is used to solve this problem, the advantage will only be more obvious and significant!
And there are some additional effects.
However, this time, before Lu Yu could say anything, the female researcher who spoke at the beginning directly stood up and refuted: “No, no, no, I think your ideas are too limited!”
“Compared with the Kalman filter, the algorithm supports the vector calculation method, which requires a lot of data and computing resources, and the requirements for the environment are much higher than that of the Kalman filter!”
“And the Kalman filter algorithm can also solve the problem of uncertain sonar interference!”
“The effect is definitely much better than the support vector machine algorithm!”
She directly said the thoughts in her mind, without any politeness.
The other group of researchers was also not polite in the slightest.
They directly retorted: “What you said is indeed true, the Kalman filter algorithm can reduce the interference caused by sonar to a certain extent, but for our micromotor system, what is the sonar problem?” ”
Its merits do exist.”
“But the drawbacks are only going to be more deadly than the support vector calculation, and you didn’t mention that at all!”
“You said that our support vector machine algorithm requires a lot of training data and computing resources, and we have no way to refute this, because it does have higher standards for resources.”
“But… What you call the Kalman filter algorithm has no shortcomings of its own?
“Yes, and it’s big!”
“Although it is not as demanding in terms of computing resources, it requires accurate models and parameters! If we build a model whose parameters are not accurate enough, its effectiveness will only be greatly reduced!
“I think it’s compared to computing resources.”
“The Kalman filter algorithm does not have any obvious advantages at all, but the support vector machine algorithm is more able to solve the problems we are currently facing!”
The female researcher did not flinch, shook her head again and said:
“I think you may have misunderstood my thinking!” The Kalman filter is like a smart glasses that adjusts the transparency and shape of the lens based on what you see and what you expect to see, allowing you to see more clearly and accurately.
“And the support vector machine is more like a powerful classifier, although the algorithm can divide different categories or regions based on the data you give it, making it easier to distinguish.”
“But for the electromagnetic interference problem we are currently facing, we need more glasses than a classifier!”
“Moreover, the Kalman filter can update the system status and predict the results in real time.”
“The support vector machine needs to train the model before it can make predictions, which is definitely more advantageous for dynamically changing systems!”
“…”
A group of people are so hyped, you don’t let me and I don’t let you, no one has the slightest intention of backing down!
They all firmly believe that the proposed algorithm is more beneficial to the development of the entire experiment.
After all, as a researcher or a relatively high-level researcher, everyone has their own arrogance, unless your ability is really far greater than him, and you can convince him, otherwise… Why do you convince him that your proposal is really better than his?
This is especially true in the face of current problems.
Whether it is the Kalman filter algorithm or the support vector calculation method, both have their own vulnerabilities and problems.
No one dares to say that the algorithm they proposed must be more advantageous.
And the degree of this problem is also very subtle, to judge which kind of problem is bigger and which one is smaller… It’s also not that obvious.
It’s not like saying that 1+1 equals several.
One answer is 2 and one answer is 3, and it is easy to determine that the correct answer is 2.
Today’s problem… I prefer whether to use Haitian oyster sauce or Lee Kum Kee’s oyster sauce for cooking tonight, and no one dares to say that the bottle of oyster sauce in their hand really presses the other party.
It’s not so easy to decide.
Although both sides of the argument this time are arguing with each other in order to make the experimental results better, they have not slept in the quarrel, and the real quarrel is a little angry!
“This comrade, I think your thinking is too limited!”
“Moreover, your thoughts are too casual!”
“You don’t think about it at all, and you just randomly think that the Kalman filter is the best choice at the moment based on a paper, which is obviously contrary to…”
“Hehe, the problem is that I don’t see that the content of your research is richer?”
“In the end, I can still get a paper that has been published to support my point of view, but you can’t even get a paper, it’s just the result of several people’s mutual research, at most, the number of people is a little advantageous over me, you say I am casual?”
“I also said you were just huddles!” The female researcher retorted.
The fatter researchers over there couldn’t help it: “You’re too much!”
“What huddle is not huddle, we are gathered together to discuss this issue, do not rise to personality attack!”
“Why have I become a personality attack again, I really want to say that the personality attack is also your predecessor personality attack!”
“Who first said I was not rigorous and that I was casual?”
“Isn’t it still you?”
The female researcher crossed her hands at her waist and looked dissatisfied.
A group of people hyped up a meeting, and the air inside was also filled with the smell of gunpowder.
Lu Yu, who was sitting in the C position, did not think too much.
He was also happy to watch the discussion intensify.
After all, although Lu Yu has the help of the system, the technology provided by the system at the beginning is more biased towards other fields, and it is important to use which algorithm to use to solve the current micromotor compatibility problem… For the time being, he can’t give a particularly good idea, or a reason, to convince everyone.
“Chief Engineer Lu, what do you think?”
After a heated discussion, the researchers found that each other’s views did not seem to convince each other, and even if they said a lot, the other party still had the initial point of view, and there was no change at all.
Then it is meaningless to continue to argue like this.
Under the deadlock between the two sides, there must be one party and a third party to break this deadlock.
Therefore, everyone also turned their eyes to Lu Yu, who was sitting in the C position, waiting for Lu Yu to give an answer.
“Hmm…” Seeing
that all eyes were focused on him, Lu Yu lowered his head slightly and thought for about a second.
“Both the Kalman filter algorithm and the support vector machine algorithm can be used as one of the alternatives.”
“As for which algorithm is better.”
“It can only be said that each has its own advantages.”
“The advantages and disadvantages of both are equally prominent, I have to say which algorithm is a better choice…” Lu
Yu bit his lip and began to think about what effect the two algorithms would have if they operated in their own way
Kalman filters can make predictions using more realistic motion models, while support vector machines need to fit models based on data
This is more advantageous for systems with clear physical laws.
In addition, the Kalman filter can better handle the error of the system and observation, r However, the support vector machine needs to worry about data preprocessing and normalization, and can also better deal with data problems with noise and outliers.
These are the advantages of the Kalman filter algorithm.
Lu Yu rubbed his temples, and his mind quickly began to think.
After half a ring, a wry smile appeared at the corner of his mouth.
If there is a saying that it is difficult for a clean official to break a housework, but sometimes, judging certain things, it feels more difficult to break than adding housework!