DRDO, PSU and Private Defence Sector News

[Okabe Rintarou]

What is AI powered smart jackets? Seems more like marketing gimmick.
Most of this AI powered system are just clever programming.
One needs to maintain 25-30 degree C inside jacket which will be best achieved with traditional PID controller.

He mentioned what was AI powered. Its just the personalization/customization algorithm for ideal temperature preference. Nothing ground breaking, they are just leveraging the buzz-word. The real innovation is in the textile. Rest all was added later on to create a marketable product. Now if they use buzzwords for marketing, its not bad is it? Or are you now saying that leveraging buzzwords is good when done by foreigners but bad when done by Indian innovators?

PID controller is supposed to adjust temperature to what exact temperature in the 25-30 range? Its not a single figure, its a multi-variable function that has ambient temperature, hydration level, relative humidity, personal preference, etc all components in it. The AI part is supposed to deal with personalizing that for optimal comfort. I don't think the jacket could have done without it.

Take another example, without AI or this textile tech, another approach that has been tried before is use of Phase-Change Material inserts. Those will not be nearly as comfortable to wear. Their approach is much more superior. Lets celebrate it.

 

[Lonewolf]

Did some calculations, diameter of vlsrsam including folded wings is 360 mm at max, while mrsam of navy has 600 mm almost without folded wing , so a uvls with 720 mm may work for us

 

[TopWatcher]

He mentioned what was AI powered. Its just the personalization/customization algorithm for ideal temperature preference. Nothing ground breaking, they are just leveraging the buzz-word. The real innovation is in the textile. Rest all was added later on to create a marketable product. Now if they use buzzwords for marketing, its not bad is it? Or are you now saying that leveraging buzzwords is good when done by foreigners but bad when done by Indian innovators?

PID controller is supposed to adjust temperature to what exact temperature in the 25-30 range? Its not a single figure, its a multi-variable function that has ambient temperature, hydration level, relative humidity, personal preference, etc all components in it. The AI part is supposed to deal with personalizing that for optimal comfort. I don't think the jacket could have done without it.

Take another example, without AI or this textile tech, another approach that has been tried before is use of Phase-Change Material inserts. Those will not be nearly as comfortable to wear. Their approach is much more superior. Lets celebrate it.

Trend going on , AI word laga jisse product high tech lagne lage.

AI is nothing but some algorithms. Aaj kal kuch jada hi hype pe hai.

 

[Okabe Rintarou]

Did some calculations, diameter of vlsrsam including folded wings is 360 mm at max, while mrsam of navy has 600 mm almost without folded wing , so a uvls with 720 mm may work for us

Doesn't 720 mm sound a lot? That is approaching BrahMos UVLS dimensions. We need a couple sizes shorter than that. What is Mk41 VLS sized at?

Trend going on , AI word laga jisse product high tech lagne lage.

AI is nothing but some algorithms. Aaj kal kuch jada hi hype pe hai.

Kuch bhi? I agree that AI is an overused buzzword, but try telling me how exactly will you solve this problem otherwise?
Mathematically AI based on Neural Networks is just a structure that can form a set of empirical relations meant to optimize a certain objective function on the fly. In this product the empirical relation for calculating optimal jacket temperature (which then becomes the target the PID controller aims for) is calculated based on the ambient dry bulb temperature, ambient relative humidity, metabolic rate of the human wearing it, perspiration level of the human wearing it, radiation incident on the jacket and personal preference on top of all such factors. All these factors I listed are functions of time with no set periodicity even though there will be some varying periodicity or some other patterns. How does your empirical relation first form and then recalibrate itself according to all these changing parameters? I don't think it can be done without Neural Networks.

I gave you the example of PCM-based jackets: after wearing that you will feel as if you were wearing two bladders (dorsal and ventral) with cold water in them. It will feel chill. Same thing happens with AC, where you need to adjust it to make it feel more comfortable. This jacket is supposed to feel comfortable without any need to turn the knob or anything like that. That is where the AI comes in: in-situ, personalized optimization.

 

[Dark Sorrow]

He mentioned what was AI powered. Its just the personalization/customization algorithm for ideal temperature preference. Nothing ground breaking, they are just leveraging the buzz-word. The real innovation is in the textile. Rest all was added later on to create a marketable product. Now if they use buzzwords for marketing, its not bad is it? Or are you now saying that leveraging buzzwords is good when done by foreigners but bad when done by Indian innovators?

PID controller is supposed to adjust temperature to what exact temperature in the 25-30 range? Its not a single figure, its a multi-variable function that has ambient temperature, hydration level, relative humidity, personal preference, etc all components in it. The AI part is supposed to deal with personalizing that for optimal comfort. I don't think the jacket could have done without it.

Take another example, without AI or this textile tech, another approach that has been tried before is use of Phase-Change Material inserts. Those will not be nearly as comfortable to wear. Their approach is much more superior. Lets celebrate it.

For all the buzz words, it is a human who will set favorable temperature as per his/her manually preference. AI won't guess what temperature a particular human requires.
When we have target set temperature then a PID controller can easily control the system in a closed loop, no need of buzz words like AI.
Their is no need for jacket to guess what is the ideal temperature rather just achieve set temperature.

 

[Okabe Rintarou]

For all the buzz words, it is a human who will set favorable temperature as per his/her manually preference. AI won't guess what temperature a particular human requires.
When we have target set temperature then a PID controller can easily control the system in a closed loop, no need of buzz words like AI.
Their is no need for jacket to guess what is the ideal temperature rather just achieve set temperature.

This is where you are wrong/lacking vision. In my previous comment above I gave the simple example of AC that you control using remote. That is what you are talking about: PID control to get to a user specified temperautre. Keep in mind that when you change AC temperature, you also use other options like removing or wearing layer of clothing, etc to optimize the temperature to your suiting. And usually you use AC inside a room shielded from elements while sitting/lying down. Not when carrying a rucksack across terrain open to elements. In the latter case, the external factors vary a lot with time making the ideal temperature that needs to be maintained to fluctuate much more frequently than the AC in the room situation.
The jacket instead aims to cut down need for user inputs after learning from user feedback over a period of time. Because that is indeed important in combat and even outside of it. Read my previous comment on the factors involved in this.
Understand the difference?

 

[TopWatcher]

Doesn't 720 mm sound a lot? That is approaching BrahMos UVLS dimensions. We need a couple sizes shorter than that. What is Mk41 VLS sized at?



Kuch bhi? I agree that AI is an overused buzzword, but try telling me how exactly will you solve this problem otherwise?
Mathematically AI based on Neural Networks is just a structure that can form a set of empirical relations meant to optimize a certain objective function on the fly. In this product the empirical relation for calculating optimal jacket temperature (which then becomes the target the PID controller aims for) is calculated based on the ambient dry bulb temperature, ambient relative humidity, metabolic rate of the human wearing it, perspiration level of the human wearing it, radiation incident on the jacket and personal preference on top of all such factors. All these factors I listed are functions of time with no set periodicity even though there will be some varying periodicity or some other patterns. How does your empirical relation first form and then recalibrate itself according to all these changing parameters? I don't think it can be done without Neural Networks.

I gave you the example of PCM-based jackets: after wearing that you will feel as if you were wearing two bladders (dorsal and ventral) with cold water in them. It will feel chill. Same thing happens with AC, where you need to adjust it to make it feel more comfortable. This jacket is supposed to feel comfortable without any need to turn the knob or anything like that. That is where the AI comes in: in-situ, personalized optimization.

I know about AI. I am trying to say its a overused word.

 

[TopWatcher]

This is where you are wrong/lacking vision. In my previous comment above I gave the simple example of AC that you control using remote. That is what you are talking about: PID control to get to a user specified temperautre. Keep in mind that when you change AC temperature, you also use other options like removing or wearing layer of clothing, etc to optimize the temperature to your suiting. And usually you use AC inside a room shielded from elements while sitting/lying down. Not when carrying a rucksack across terrain open to elements. In the latter case, the external factors vary a lot with time making the ideal temperature that needs to be maintained to fluctuate much more frequently than the AC in the room situation.
The jacket instead aims to cut down need for user inputs after learning from user feedback over a period of time. Because that is indeed important in combat and even outside of it. Read my previous comment on the factors involved in this.
Understand the difference?

Better call it ML not AI there is a huge difference b/w AI & ML.

 

[notaname]

What is AI powered smart jackets? Seems more like marketing gimmick.
Most of this AI powered system are just clever programming.
One needs to maintain 25-30 degree C inside jacket which will be best achieved with traditional PID controller.

Trend going on , AI word laga jisse product high tech lagne lage.

AI is nothing but some algorithms. Aaj kal kuch jada hi hype pe hai.

For all the buzz words, it is a human who will set favorable temperature as per his/her manually preference. AI won't guess what temperature a particular human requires.
When we have target set temperature then a PID controller can easily control the system in a closed loop, no need of buzz words like AI.
Their is no need for jacket to guess what is the ideal temperature rather just achieve set temperature.

Quotes from Article what exactly it does.

After some data feeding the system will automatically adjust to different temperatures accordingly as per the needs, which you wouldn't be able to do in traditional programming without adding in Probability and data based algo's which can be done using AI and ML.

When we initially began work, we wanted to base the model on the concept that the user could regulate heat generation at his or her convenience,” says Khusboo. “But since it was to be specifically designed for the armed forces, making them fiddle with the regulators and switches would be an unnecessary burden to impose on them. They shouldn’t have to focus on these trivial aspects when their prime duty is so critical in itself.”

The next step was to focus on coming up with a mechanism that would ensure the jacket maintains a comfortable range of temperature on an auto mode. However, as the duo would soon realise, “comfort” was subjective — while one personnel might want a certain temperature to keep warm, for another, that may be uncomfortable.

“The software would learn the wearer’s personal preferences in different settings within a short period of time and then begin to adjust to their comfortable temperature.”

 

[TopWatcher]

Quotes from Article what exactly it does.

After some data feeding the system will automatically adjust to different temperatures accordingly as per the needs, which you wouldn't be able to do in traditional programming without adding in Probability and data based algo's which can be done using AI and ML.

Better call it ML not AI.

 

[Okabe Rintarou]

Better call it ML not AI there is a huge difference b/w AI & ML.

Better call it ML not AI.

No. Better to call it AI even if its just a Neural Network. Its called marketing. Our businesses need to use it where possible. Everybody does it. No need to adopt stricter guidelines for our own products while our forces and public consume foreign wares with similar marketing hyperbole.

 

[notaname]

Better call it ML not AI.

ML is subset of AI, and after some data processing the system will go from data interpretation to data prediction and use it properly, so this can be termed as AI, and for marketing purposes the end users only know AI not ML so there is that. If you want to go in technicalities we can add in Data Analytics also, or Pure math functions with some catchy algorithm names.

Even if AI is overused word, there is no harm in using that as not everyone knows what exactly AI does.

 

[TopWatcher]

No. Better to call it AI even if its just a Neural Network. Its called marketing. Our businesses need to use it where possible. Everybody does it. No need to adopt stricter guidelines for our own products while our forces and public consume foreign wares with similar marketing hyperbole.

That what i meant to say

 

[TopWatcher]

ML is subset of AI, and after some data processing the system will go from data interpretation to data prediction and use it properly so, this can be termed as AI, and for marketing purposes the end users only know AI not ML so there is that, if you want to go in technicalities we can add in Data Analytics also here, or Pure math functions with some catchy algorithm names.

Even if AI is overused word, there is no harm in using that as no everyone knows what exactly AI does or how it is attained.

Yeah, i loved this field data analytics , ML & AI.

 

[Okabe Rintarou]

That what i meant to say

I do feel just how overused/misused this term is:-

 

[notaname]

I do feel just how overused/misused this term is:-

View attachment 172899

Even in majority of Indian collages, Professers or Visiting experts demo ML and term it AI.

 

[flanker99]

Posting this here as its kind of the same concept that @porky_kicker talked about a hybrid air water propulsion system being developed by drdo....

- YouTube

Enjoy the videos and music that you love, upload original content and share it all with friends, family and the world on YouTube.

youtube.com

 

[Lonewolf]

Doesn't 720 mm sound a lot? That is approaching BrahMos UVLS dimensions. We need a couple sizes shorter than that. What is Mk41 VLS sized at?

Similar dimensions

 

[flanker99]

Indian Army Prepares To Procure Hundreds Of Armed Swarm Drones And Loitering Munitions

Various types of swarm drones are being developed for the army. IAF is also seeking swarm drones for various roles, including FOD detection in airfields.

www.overtdefense.com

 

[Photon]

From my personal collection (DAE 2018 brochure)