Indian Navy Developments & Discussions

[FalconSlayers]

View: https://x.com/OfficialGRSE/status/1849777699427950986

 

[darkflame]

so with the commissioning of last kalvari class submarine by end of this year and 3rd arihant class submarine Aridhaman early next year we will end up having 20 submarines in total

 

[shade2]

so with the commissioning of last kalvari class submarine by end of this year and 3rd arihant class submarine Aridhaman early next year we will end up having 20 submarines in total

Our seas should be aircraft carrier mukt.

 

[Azaad]

so with the commissioning of last kalvari class submarine by end of this year and 3rd arihant class submarine Aridhaman early next year we will end up having 20 submarines in total

Another way of looking at things is in this decade we may well end up inducting only 1 Akula class SSN from Russia if at all .

In the meanwhile , all the 6 Scorpenes will be due for MLU with India make AIP from next year onwards. That means each submarine will be out of action for a good 3-4 years.

By the middle to end of the next decade all the HDW & Kilo class submarines will be retired. As it is their performance now is sub optimal give their age . By the next decade they'd barely move outside the harbour. Their replacements be it the additional 3 Scorpenes or thru the Project 75 I have yet to see the contracts being signed.

Further we'd see 2 SSNs vide Project 77 & possibly 1-2 S-5 class late next decade. So it's a mixed bag.

 

[darkflame]

Another way of looking at things is in this decade we may well end up inducting only 1 Akula class SSN from Russia if at all .

In the meanwhile , all the 6 Scorpenes will be due for MLU with India make AIP from next year onwards. That means each submarine will be out of action for a good 3-4 years.

By the middle to end of the next decade all the HDW & Kilo class submarines will be retired. As it is their performance now is sub optimal give their age . By the next decade they'd barely move outside the harbour. Their replacements be it the additional 3 Scorpenes or thru the Project 75 I have yet to see the contracts being signed.

Further we'd see 2 SSNs vide Project 77 & possibly 1-2 S-5 class late next decade. So it's a mixed bag.

i think kilos and type 209s inducted in 90s most likely will get a life extension refit and may continue to operate till 2040s, agree with every thing else though.

 

[Ayan Barat]

i think kilos and type 209s inducted in 90s most likely will get a life extension refit and may continue to operate till 2040s, agree with every thing else though.

"Continue to operate" is a vague parameter as you can continue to operate MiG-21 today but the question is how good it'll be?

What he meant by age is not the fact that whether it was inducted in 1980 or 1990 but more with the fact that it was designed in 1968. Midlife upgrades can definitely extend the service life of an equipment by improving the electronics or adding better communication systems but it can't deal with things like detectability or engine noise.

 

[SlayingHell]

Another way of looking at things is in this decade we may well end up inducting only 1 Akula class SSN from Russia if at all .

In the meanwhile , all the 6 Scorpenes will be due for MLU with India make AIP from next year onwards. That means each submarine will be out of action for a good 3-4 years.

By the middle to end of the next decade all the HDW & Kilo class submarines will be retired. As it is their performance now is sub optimal give their age . By the next decade they'd barely move outside the harbour. Their replacements be it the additional 3 Scorpenes or thru the Project 75 I have yet to see the contracts being signed.

Further we'd see 2 SSNs vide Project 77 & possibly 1-2 S-5 class late next decade. So it's a mixed bag.

They should fully pivot towards a combination of SSNs and XLUUVs (100t-500t), as SSKs don’t offer any meaningful advantage for their mission needs. For tasks like coastal patrol and choke-point operations, XLUUVs are a superior choice. After all, a human in the loop can only respond to what their sensors detect; there’s nothing inherently unique to having a human onboard that can’t be programmed into these systems. Missions conducted by SSKs are effectively high-risk ventures: if the first volley isn’t successful, the vessel becomes a sitting duck.

Each SSK is costing around $1 billion, and we lack the manufacturing efficiency that China, Korea, or Japan have in building them cost-effectively. Investing $1 billion per SSK doesn’t make financial sense when compared to the returns on a $3 billion SSN, which offers far greater value for the investment.

At this point, it would be more practical to complete the acquisition of three more Scorpenes and conclude the SSK program there. Cancel the P75I and P76 projects, and instead focus on establishing an SSN production line similar to the U.S. model. With serialized production and batch-wise improvements, we could aim to produce one SSN per year. Given an SSN lifespan of roughly 32 years, this would allow us to maintain an equilibrium fleet of 32 SSNs.

 

[Ayan Barat]

XLUUVs

Yup but first clear few of my doubts

1. Almost all the frequencies gets absorbed by water except for VLF and ELF and by the very definition of frequencies, these are 10-100km high wavelength that needs ridiculous antenna systems. The most common being the trailing wire antennas on doomsday planes that are used only in as the name suggests, doomsday where the plane flies as close to the submarine as possible and transmit as short of a message as possible like "N China".
So how are you going to manage the control of that UUV?

2. Even in the previous case, the communication is very basic with sometimes being as simple as Morse code but more importantly; it's just one way. A submarine is never supposed to talk back unless it's sinking.
So how will you manage the transfer of high amount of high fidelity data from all the sensors of the submarine to the control station and then the commands from there with as little lag as possible?

3. Currently submarines have three vulnerabilities; they can be detected by SONARs, thier noise can be picked up by hydrophones or their conning towers may get caught by SARs when they surface. But now you're adding a fourth vulnerability to it; the RF.
How are you going to deal with enemy ECCMs picking up the comlink between the UUV and the control or even worse; electronic warfare?

there’s nothing inherently unique to having a human onboard that can’t be programmed into these systems.

A 20 man crew can lurk with their engines turned off and SONARs switched to hydrophone mode for 10 days. No radio communication, no ping, no engine noise...just listening to all the sounds around them. Based on the sounds and the immense experience of the crew, they can accurately identify a target and engage it...all without a single noise other than the slight disturbance by a torpedo tube getting flooded.

I guess this can be termed as something inherently unique to having crew onboard. If even it's not enough then on one hand you'd need to transmit data from hundreds of different sensors back to the control station just to know whether it's working or not and on the other two guys can go to the engine room to change a faulty bearing.

 

[Azaad]

Yup but first clear few of my doubts

1. Almost all the frequencies gets absorbed by water except for VLF and ELF and by the very definition of frequencies, these are 10-100km high wavelength that needs ridiculous antenna systems. The most common being the trailing wire antennas on doomsday planes that are used only in as the name suggests, doomsday where the plane flies as close to the submarine as possible and transmit as short of a message as possible like "N China".
So how are you going to manage the control of that UUV?

2. Even in the previous case, the communication is very basic with sometimes being as simple as Morse code but more importantly; it's just one way. A submarine is never supposed to talk back unless it's sinking.
So how will you manage the transfer of high amount of high fidelity data from all the sensors of the submarine to the control station and then the commands from there with as little lag as possible?

3. Currently submarines have three vulnerabilities; they can be detected by SONARs, thier noise can be picked up by hydrophones or their conning towers may get caught by SARs when they surface. But now you're adding a fourth vulnerability to it; the RF.
How are you going to deal with enemy ECCMs picking up the comlink between the UUV and the control or even worse; electronic warfare?

A 20 man crew can lurk with their engines turned off and SONARs switched to hydrophone mode for 10 days. No radio communication, no ping, no engine noise...just listening to all the sounds around them. Based on the sounds and the immense experience of the crew, they can accurately identify a target and engage it...all without a single noise other than the slight disturbance by a torpedo tube getting flooded.

I guess this can be termed as something inherently unique to having crew onboard. If even it's not enough then on one hand you'd need to transmit data from hundreds of different sensors back to the control station just to know whether it's working or not and on the other two guys can go to the engine room to change a faulty bearing.

Since you're the resident tech expert here why don't you elaborate on the solution after diagnosis of the problem. He's run with the baton in a relay , perhaps you could run with the baton upto the next runner who in turn would add more to the learnings here & maybe lay people like us could learn a thing or two .

ULF & VLF communication systems are being used to communicate with SSBNs . Why can't they be utilised to communicate with UUVs. Perhaps we need to look at the mother ship concept where we have a manned submarine connected to a number of UUVs controlling their movement like what is planned with the MUMT in air warfare.

Or perhaps communication systems for these kind of operations aren't that well-developed & are in the process of being developed for such usage & up until then we'd be in experimental mode for the idea itself is solid given the risk factor . I'd extend the same to tank warfare , hptrs in CAS role etc . Frankly in the next 3-4 decades we should be there with these systems.

 

[SlayingHell]

A 20 man crew can lurk with their engines turned off and SONARs switched to hydrophone mode for 10 days. No radio communication, no ping, no engine noise...just listening to all the sounds around them. Based on the sounds and the immense experience of the crew, they can accurately identify a target and engage it...all without a single noise other than the slight disturbance by a torpedo tube getting flooded.

Are you really suggesting that a human operator on board can perform the same level of processing, identification, and engagement as a computer? The U.S. Navy possesses an advanced Underwater Threat Identification Library that is capable of analyzing complex data, including discerning the maintenance state of Russian submarines. This kind of sophisticated data interpretation is beyond human capability; a person could spend a lifetime and still not be able to process that information as efficiently as an algorithm.

When it comes to decision-making in the field, a human operator typically reacts to the data presented on their screens. They do not possess the capacity to analyze and interpret that vast array of information independently.

Regarding communication, how can a human operator onboard an SSK know when to engage? This situation is not fundamentally different for UUVs, which can be designed to receive mission parameters and act based on pre-established rules or real-time data, whether they communicate periodically or use RF. If an SSK crew is given explicit instructions before a mission, the same concept applies to UUVs.

If an SSK relies on RF communications for engagement decisions, then UUVs can be programmed to operate under similar parameters. Ultimately, I struggle to understand how human decision-making goes beyond a simple binary choice of “yes” or “no” for engagement. If there are submariners out there who disagree, I welcome their input, but my focus remains solely on SSKs and not on SSNs or SSBNs.

Note all above of these arguments are for SSKs only.

 

[Ayan Barat]

the resident tech expert

Never claimed to be. The only thing I'm is...

I'm bit retarded

He's run with the baton in a relay , perhaps you could run with the baton upto the next runner who in turn would add more to the learnings here & maybe lay people like us could learn a thing or two .

Relay race?
Naaah...I see myself more like the guy who's job is to increase the height of the bar after each long jump or pole vault to stimulate others to jump as high as possible

ULF & VLF communication systems are being used to communicate with SSBNs . Why can't they be utilised to communicate with UUVs.

Not communicate, just send one way messages
As for why can't they be utilised look at the closest thing we have to a mothership and an UUV; a submarine and a torpedo. Now ask the question that why almost all the torpedoes use fibre optics for guidance instead of radio frequencies.

If you give an SSK crew instructions before it goes to mission to engage enemies, then same is the case for UUVs.

For this long I was thinking the model you're proposing was that of an "UUV sends SONAR data > control centre sends back firing command > it fires" something we've been doing with UCAVs. Missed the fact that you're taking about a completely autonomous system. Sorry for that

If you're talking about an AI enabled completely autonomous UUV then the answer is a clear no. We're yet to develop an autonomous system reliable enough to give it the power to target a near peer adversary's naval fleet.
And even if somehow any navy on this planet comes up with this kind of system then its parliament would shoot the plan down.

 

[SlayingHell]

Never claimed to be. The only thing I'm is...


Relay race?
Naaah...I see myself more like the guy who's job is to increase the height of the bar after each long jump or pole vault to stimulate others to jump as high as possible

Not communicate, just send one way messages
As for why can't they be utilised look at the closest thing we have to a mothership and an UUV; a submarine and a torpedo. Now ask the question that why almost all the torpedoes use fibre optics for guidance instead of radio frequencies.

For this long I was thinking the model you're proposing was that of an "UUV sends SONAR data > control centre sends back firing command > it fires" something we've been doing with UCAVs. Missed the fact that you're taking about a completely autonomous system. Sorry for that

If you're talking about an AI enabled completely autonomous UUV then the answer is a clear no. We're yet to develop an autonomous system reliable enough to give it the power to target a near peer adversary's naval fleet.
And even if somehow any navy on this planet comes up with this kind of system then its parliament would should the plan down.

Why is AI necessary in this context? Consider what a fly-by-wire (FBW) system or the Soviet Buran spacecraft is—autonomous systems that operate without requiring AI to function effectively. Autonomy does not equate to artificial intelligence.

Let’s examine two scenarios:

**Scenario 1:** An SSK crew conducts their mission without any communication. Before deployment, they receive explicit instructions to engage enemy submarines. In this case, an XLUUV can be programmed with the same instructions before it heads to its designated area. Once in position, if the XLUUV’s onboard computer detects a target that matches its threat library, it can engage that target. Humans, in this scenario, are performing a similar function—they, too, are making a binary decision based on the data they have: a simple “yes” or “no” to engage.

**Scenario 2:** In this case, the SSK crew sends data back to command for engagement decisions. Again, an XLUUV could be designed to perform the same task, relaying data and receiving commands without human intervention.

It’s important to note that automation has been in existence long before AI came onto the scene.

 

[Ayan Barat]

Why is AI necessary in this context? Consider what a fly-by-wire (FBW) system or the Soviet Buran spacecraft is—autonomous systems that operate without requiring AI to function effectively. Autonomy does not equate to artificial intelligence.

Let’s examine two scenarios:

**Scenario 1:** An SSK crew conducts their mission without any communication. Before deployment, they receive explicit instructions to engage enemy submarines. In this case, an XLUUV can be programmed with the same instructions before it heads to its designated area. Once in position, if the XLUUV’s onboard computer detects a target that matches its threat library, it can engage that target. Humans, in this scenario, are performing a similar function—they, too, are making a binary decision based on the data they have: a simple “yes” or “no” to engage.

**Scenario 2:** In this case, the SSK crew sends data back to command for engagement decisions. Again, an XLUUV could be designed to perform the same task, relaying data and receiving commands without human intervention.

It’s important to note that automation has been in existence long before AI came onto the scene.

Let's assume everything happens successfully.
As you explicitly mentioned A NON ARTIFICIAL INTELLIGENT autonomous UUV with algorithms capable enough to be a true hunter-killer on its own is developed.

The civilians in charge of the military would never allow it to have a system that can take the decision to engage the target on its own without a human-in-loop.

As simple as it is.

 

[SlayingHell]

Let's assume everything happens successfully.
As you explicitly mentioned A NON ARTIFICIAL INTELLIGENT autonomous UUV with algorithms capable enough to be a true hunter-killer on its own is developed.

The civilians in charge of the military would never allow it to have a system that can take the decision to engage the target on its own without a human-in-loop.

As simple as it is.

Yeah it boils down to leadership not comfortable in such. But they are comfortable in high risky missions of SSK. And also splurging 1-Billion plus on a SSK, at a price point its a wasteful spending.

SSKs are okay as long as they don't cost beyond certain price like 500M or 750M dollars and anything beyond that is absolute wastage of money better spent on improvising SSN program. It is the very reason all SSN capable players like USA, UK, France don't do SSKs. Russkies did Kilo class because their coastline is vast and they couldn't build on level of West. Chinese did SSK because only recently they upped their SSN game and they can build it cheap.

 

[Binayak95]

i think kilos and type 209s inducted in 90s most likely will get a life extension refit and may continue to operate till 2040s, agree with every thing else though.

the Type 209s and the last trench of Kilos have already been given life extension programs, and mid life upgrades, and everything else.
They are due for scrapping. The GoI should have really taken the additional 3 scorpene option 3 years ago.

 

[Binayak95]

They should fully pivot towards a combination of SSNs and XLUUVs (100t-500t), as SSKs don’t offer any meaningful advantage for their mission needs. For tasks like coastal patrol and choke-point operations, XLUUVs are a superior choice. After all, a human in the loop can only respond to what their sensors detect; there’s nothing inherently unique to having a human onboard that can’t be programmed into these systems. Missions conducted by SSKs are effectively high-risk ventures: if the first volley isn’t successful, the vessel becomes a sitting duck.

Each SSK is costing around $1 billion, and we lack the manufacturing efficiency that China, Korea, or Japan have in building them cost-effectively. Investing $1 billion per SSK doesn’t make financial sense when compared to the returns on a $3 billion SSN, which offers far greater value for the investment.

At this point, it would be more practical to complete the acquisition of three more Scorpenes and conclude the SSK program there. Cancel the P75I and P76 projects, and instead focus on establishing an SSN production line similar to the U.S. model. With serialized production and batch-wise improvements, we could aim to produce one SSN per year. Given an SSN lifespan of roughly 32 years, this would allow us to maintain an equilibrium fleet of 32 SSNs.

Aint no one ever replacing SSK with XLUUV, you cant talk to the bloody things and no commodore worth his salt trusts his own commander, let alone a fucking autonomous ship killer that is one mis - classification away from kamakazing INS Vikrant.

 

[Binayak95]

Yeah it boils down to leadership not comfortable in such. But they are comfortable in high risky missions of SSK. And also splurging 1-Billion plus on a SSK, at a price point its a wasteful spending.

SSKs are okay as long as they don't cost beyond certain price like 500M or 750M dollars and anything beyond that is absolute wastage of money better spent on improvising SSN program. It is the very reason all SSN capable players like USA, UK, France don't do SSKs. Russkies did Kilo class because their coastline is vast and they couldn't build on level of West. Chinese did SSK because only recently they upped their SSN game and they can build it cheap.

USA and UK and France dont do SSKs because the fuckers dont have the manpower to crew the bloody things. SSNs are absolutely disgusting waste of money for standard duty operations.

France and RN are INCAPABLE of doing your normal submarine mission profile of say, attacking the enemy's merchant shipping, They're dedicated SSBN escorts.

For everything else, be it interdiction of enemy merchantmen, for sulking around and gathering SIGINT and laying mines, and inserting SF - SSKs are the best thing.

SSNs are noisier, more expensive to build and operate and the crew qualification requirements are off the charts. Its a stupid bet for non escort SSBN missions.

 

[Binayak95]

Why is AI necessary in this context? Consider what a fly-by-wire (FBW) system or the Soviet Buran spacecraft is—autonomous systems that operate without requiring AI to function effectively. Autonomy does not equate to artificial intelligence.

Let’s examine two scenarios:

**Scenario 1:** An SSK crew conducts their mission without any communication. Before deployment, they receive explicit instructions to engage enemy submarines. In this case, an XLUUV can be programmed with the same instructions before it heads to its designated area. Once in position, if the XLUUV’s onboard computer detects a target that matches its threat library, it can engage that target. Humans, in this scenario, are performing a similar function—they, too, are making a binary decision based on the data they have: a simple “yes” or “no” to engage.

**Scenario 2:** In this case, the SSK crew sends data back to command for engagement decisions. Again, an XLUUV could be designed to perform the same task, relaying data and receiving commands without human intervention.

It’s important to note that automation has been in existence long before AI came onto the scene.

a highly trained, decades experience sonar operator can at best guesstimate whether or not a contact on the sonar is a bloody type 54D or a Burke.

If you think ML can classify with better accuracy, newsflash mate. It aint happening. We simply dont even have the dataset to even approach a avg prediction accuracy of 40%.

And relaying data, from an XLUUV?

If your XLUUV is carrying that kinda communication power, it aint stealthy anymore, the nearest MR bird is tango 10 out of dropping a homing torpedo.

 

[Azaad]

https://swarajyamag.com/defence/explained-whats-holding-up-indian-navys-project-to-build-submarines-that-can-stay-underwater-longer

 

[Ayan Barat]

@SlayingHell

Researchers say an AI-powered transcription tool used in hospitals invents things no one ever said

Whispher is a popular transcription tool powered by artificial intelligence but it has a major flaw. It makes things up that were never said.

apnews.com

This is the biggest problem...