Chapter 825 Testing, Target Feature Extraction

Of course, using ballistic missiles to attack aircraft carriers is definitely not that simple. It can be solved by creating a target classification algorithm.

First of all, the prerequisite for “target classification” is to detect the target.

But how to make the ballistic missile overcome the black barrier and use the radar normally during the re-entry into the atmosphere is a huge problem...

Reducing the reentry speed to below 5 times the speed of sound can avoid the blackout at its source. However, low-speed ballistic targets are the easiest targets to deal with for air defense systems, which goes against the original purpose of using ballistic missiles against ships.

If the radar is turned on after passing the blackout area, the reaction time will be too short and there will not be enough time for the missile to adjust its attitude.

Secondly, in order to find the location of an aircraft carrier battle group in the vast ocean before launching an attack, a complete global ocean monitoring system is needed.

At this moment, China has just launched its first land-transmission remote sensing satellite.

The second and third are Resource 01 02 and which are about to be launched.

This was also the original reason why Chang Haonan rushed out to conduct technical testing before the Spring Festival.

As for ocean remote sensing satellites...

Sorry, there aren't any yet.

In short, at least in 2000, the idea of ​​developing an anti-ship ballistic missile was still just a thought.

Most people don't even dare to think about it.

On the contrary, Wang Xiaomo's ideas are more realistic.

In particular, Chang Haonan had previously considered designing an anti-submarine patrol aircraft based on the C808 platform with comfort and range exceeding or not inferior to the P3C.

In the current context where the main tone of confrontation is to fight without breaking, the importance of this type of equipment is at least no less than that of conventional main combat equipment such as tactical aircraft.

Thinking of this, Chang Haonan immediately turned to look at Liang Shaoxiu, who was unable to get involved in the topic.

"Engineer Liang, our KJ-200 doesn't have hardware to shield low-altitude targets, right?"

Since the main function considered during the development of early warning aircraft is air detection, there are some models that directly limit the longitudinal pitch angle of electromagnetic waves in hardware, in order to solve the interference caused by ground echoes in the simplest and crudest way.

What's more, a shielding layer will be designed under the antenna to directly block the downward electromagnetic waves with large depression angles.

This thing definitely wasn't included when Chang Haonan originally made the plan, but there's no guarantee whether it might have been added in the past two years.

"No, we just set a downward viewing angle limit in the radar background program."

Liang Shaoxiu replied:

"Because it is still a prototype, this limitation can be lifted through special commands. If you have a new target recognition program, you can just modify the software here."

Since pilots and ground crew are not professional aircraft designers, when mass-produced aircraft come into their hands, in order to prevent people from tampering with them, they lock the software programs and only leave some preset input and output ports.

However, prototypes or technology verification machines will have a state similar to "developer mode", allowing engineers to make modifications more conveniently.

"Don't worry, there is no need to make major changes to the current procedures."

Chang Haonan turned around and looked at the staff who followed him.

The latter immediately understood and took out a password box he carried with him.

After opening it, there were several hard drives inside.

"I have encapsulated the target feature extraction and analysis functions into a module that can be plugged into the existing data processing code."

Having said that, the operating computer on the early warning aircraft is not a home computer after all. All you need to do is create an installation file in .msi format and wait for it to complete automatically.

Windows XP, which was later widely used on military terminals, was still a trendy thing that could only be installed on the latest PCs.

It has extremely high hardware requirements and is unstable.

It's almost the same as win2023 in 11.

And it hasn’t been officially released yet.

Therefore, it took Chang Haonan nearly a week to completely complete the program update and necessary ground testing.

It even took him longer than it had taken him to write programs before.

Of course, Liang Shaoxiu and his team were not idle during these days.

Instead, he helped coordinate with the naval leaders and borrowed a batch of fake targets for training.

It’s not something fancy like a self-propelled decoy.

Instead, they simply use primitive methods to install different styles of models or corner reflectors on the buoys.

Used to simulate targets of various sizes, from sampans to destroyers, at low cost in non-live-fire training.

It was a method that the army had to use to endure the period in the early years.

The advantage is that it can be deployed with just a transport boat or tugboat, and can be recovered intact after use.

But later I found that the effect was actually quite good.

However, because it is a buoy, it certainly cannot be deployed in places like the high seas and oceans, and can only function in relatively shallow locations.

……

The last day before New Year’s Eve.

"General Chang, the comrades from the North Sea Fleet sent back a message saying that they have already set up all 35 pre-set fake targets. In addition, the three warships 107, 110, and 519 that are currently training in nearby waters will also cooperate with us and serve as real ship targets to compare with the signal characteristics of the corner reflectors."

Liang Shaoxiu hung up the phone and said to Chang Haonan who was not far away.

The latter was at the tower's floor-to-ceiling window, watching the KJ-200 drive from the apron onto the taxiway:

"Notify flight 20001 and prepare for takeoff."

If it is reasonable, then as Chang Haonan is the deputy chief designer, Liang Shaoxiu should be considered his leader in the KJ-200 project.

However, the content of this test actually has nothing to do with the early warning aircraft itself, and the two sides are closer to cooperation.

The reason why it is carried out within the framework of KJ-200 is simply for convenience.

Moreover, since the end of 96 when the wing icing problem of the Y-8 platform was solved, Liang Shaoxiu has actually been learning from Chang Haonan.

Just like Lin Shikuan of Institute 601 back then, he can be considered a half-apprentice of Chang Haonan.

Therefore, there is a strange scene in front of us where the chief engineer reports the situation to the deputy chief engineer, and then the deputy chief engineer gives instructions to the chief engineer.

A few minutes later, the KJ-200 took off from the end of the runway, leaving four trails of light black smoke, and gradually disappeared from everyone's sight.

"Boss Chang, Comrade Shao Xiu has complained to me before, saying that the performance of the turboprop 6 is still a bit lagging behind..."

Wang Xiaomo looked at the black trail that had not yet completely dissipated in the air, and said to Chang Haonan beside him:

"I heard that when your Aviation Power Group was established last year, you mentioned a plan to develop a turboprop engine?"

Obviously, this is a hint...or an indication, of wanting a turboprop engine with better performance.

In other words, a flying platform with better performance.

If Liang Shaoxiu said this, it would be difficult to strike the right balance between suggestion and dissatisfaction.

But Wang Xiaomo has nothing to do with the aviation system, so he doesn't have any concerns in this regard.

"Indeed."

Chang Haonan nodded:

"But the design power of that machine is only 1600 kW. It is intended for 10-ton helicopters or 7-ton turboprops like the Y-20. They are not in the same class."

"As for the 4000kw turboprop... I won't say anything fancy, it really doesn't have a high priority at the moment."

"We do have the R&D capabilities, but we have too few technicians to handle too many models at the same time. We can only focus on the most urgent ones first..."

"..."

When the KJ-200 landed again, it was almost three hours later.

According to Chang Haonan's test requirements, the aircraft must conduct four rounds of searches for the target from different angles, that is, two zigzag routes that are roughly perpendicular to each other.

For the KJ-200, whose range is already somewhat limited, this is even the longest flight since the prototype's first flight...

While the ground crew rushed over to perform maintenance on the aircraft, the onboard engineer had already taken down the recorded radar data and handed it over to Chang Haonan.

Of course, there is no supercomputer available at the Xingcheng base.

Fortunately, if you just want to check the results or make simple parameter adjustments, you can use an ordinary computer.

But if the algorithm needs to be improved later, we will have to go back to Beijing to do it.

"We marked targets 1-39 in the order in which we first completed the search."

The radar operator who was in charge of following up the test flight was a naval lieutenant colonel named Su Changyun. He was sitting behind Chang Haonan and Wang Xiaomo, explaining the details of the flight:

"Originally, there should have been 35 decoys plus three ships, a total of 3, but we found an illegal fishing boat in the target waters, so there was one more."

"Then this fishing boat finally..."

Although this matter has nothing to do with the test, Chang Haonan still couldn't suppress his curiosity.

"We contacted the Dalian Ship 110 through the North Sea Fleet Command and finally caught them. We found out that they were a group of illegal immigrants."

"Okay... that's good."

Wang Xiaomo next to him showed a somewhat helpless smile:

"To be the first victory in our country's history under the command of an early warning aircraft, to some extent it is recorded in the history books..."

"It also provides us with a typical signal characteristic of a fishing vessel."

Chang Haonan added:

"In the future, it can be added to the whitelist of anti-ship weapons, or at least the aiming coefficient can be put at the end to avoid accidental hits..."

The other people here naturally didn't know about the joke about the Hsiung Feng III attacking the fishing boat, so they just thought he was preparing for a rainy day.

After this little episode, everyone turned their attention back to the radar data.

There are two groups of data in total, one is the original data and the other is processed by the recognition algorithm.

In addition to being used for comparison, the difference between the two sets of data can also be used as a basis for judging the nature of the target.

Since the radar display is not specially configured, the result outputted on it is a dot.

But in the background operation log, the difference can be clearly seen.

The most intuitive thing is that the latter has richer or more distinct data levels than the former.

"There are two targets, 1 and 2, in four directions, and two sets of data for each. The first set was hidden in the clutter on the sea surface and was not seen. The second set was seen but very faintly. It should be a simulated periscope..."

"The RCS of targets 3 to 16 are quite large, but it was exposed in the second set of data. The signal characteristics of targets 36, 37, and 38 are completely different... Considering the difference in quantity, the first dozen should be different large corner reflectors, and the last three should be the three real warships..."

"17 and 18, the signal characteristics detected in the four directions are almost the same. They should be small corner reflectors used to simulate speedboats or patrol boats..."

"..."

Every time Chang Haonan finished a sentence, Su Changyun would compare it with the "standard answer" sent by the fleet.

In the end, except for the wrong judgment of "reef" and "small boat", everything else was correct.

For a first test, the results are pretty good -

After all, radar and computers don't know what humans define as a target; they are just responsible for figuring out the difference.

As for the inductive analysis of signals with different characteristics, that needs to be calibrated slowly later.

For example, targets with feature A are large and medium-sized surface ships, targets with feature B are ordinary bulk carriers, and targets with feature C are container ships...

In fact, this is exactly the fundamental purpose of Chang Haonan coming here to conduct the test.

Although the situation of early warning aircraft cannot be directly applied to remote sensing satellites, the methods of feature extraction and selection of calibration anchor points are similar.

“Although it’s not perfect yet, it does have great potential.”

After looking at all the results, Wang Xiaomo came to a conclusion almost immediately:

"Especially in terms of decoy identification, it is absolutely accurate at a glance, but unfortunately the storage and computing capacity required are too large to be easily placed on a fighter's radar."

"That may not be the case."

Chang Haonan was already adjusting the parameters and calibration methods, but he stopped when he heard the other party's words:

"The main function of fighter radar is air search. The background is relatively clean and the target features are more monotonous. I can write a simplified version for the 1473 and 1493 series radars..."

"!"

Wang Xiaomu was just saying it casually just now, but he didn't expect that Chang Haonan could actually come up with a feasible plan.

He stood up quickly:

"Then I'll contact the 14th Institute. These radars, oh, and the 346, haven't been finalized yet. We can take this opportunity to work on them together..."