I am interested to understanding the gap between "the limits imposed by the nature" and our state of the art technology. I suppose understanding this will give some clues for future breakthroughs. Comments on this will be interesting.
Thk-you ever so much Dr Emil Björnson. I just watched a few of your older videos & I finally got a handle on the MIMO big-picture. Of course the details still make my head explode (very messy ;). But as long as I keep-at-it, I'm sure that'll happen much less often. Of course I'll download your book too, thks-again. And I thought Sweden was only well-known for IKEA & ABBA. I wonder what else you Swedishs are great at?
It was very interesting to learn about holographic MIMO. In millimeter wave imaging, the scattered field at the image plane is reconstructed by Fourier transform to obtain a snapshot of the object plane with the depth information included. I am curious how this kind of reconstruction can be performed in the context of wireless communication?
I was really enjoying Your discussion. Let me dare ask You following question: You referred a lot to Aperture. I am seasoned Telco consultant and passionate photographer as well - Can Aperture in antennas be used in antennas in same way is in camera lenses? In photography its not fixed and static, rather Aperture is dynamic, where via adapting blades You are manipulating light thruput via lens according Your needs. Do You see any benefit doing same for antennas apertures?
There are some similarities between these concept. The aperture of an antenna array is basically its physical size and we are usually using all the antennas to maximize the signal strength. We are not changing the aperture. But suppose we would turn off the outer antennas in the array to reduce the aperture. This will lead to wider beams and larger beam depth (if we operate in the radiative near field), at the cost of a power reduction. This is similar to what happens in a camera: you reduce the aperture to get a larger depth, but pay the price in terms of less light. There can be a benefit of reducing the array aperture when you want the beam to cover the entire coverage area. But there are more clever ways to do the same thing using the entire array to transmit a wide beam. It is like cutting the array into many smaller pieces, operate each with a smaller aperture and then combine them to not lose any power.
Thanks you a lot for this amazing rich podcast! I have one question though! when the discussion was about the usage of massive MIMO in Thz bands tom wasn't so positive about it. So one can say that massive MIMO is not necessary to be used within the very high frequency bands?
I believe that Tom have spatial multiplexing of many user devices in mind when he talks about Massive MIMO. In THz bands, you will for sure need many antenna elements to get a decent link budget, but perhaps not use them for spatial multiplexing since the range will be short so the chance of having many users is fairly small.
Strictly speaking, the delay spread isn't frequency dependent. However, when you change the carrier frequency and the array structure, the relative strength of the different paths might change and effectively reduce the delay spread. That said, I don't think it will be inversely proportional to the carrier frequency. The point that was made in the podcast is that the coherence time is inversely proportional to the carrier frequency, which will reduce the size of a coherence block, unless the delay spread also is inversely proportional (which it probably isn't)
Sir, so in mmWave technology channel estimation is a unrealizable dream? because channel coherence time decrease by 10 microsecond. kindly reply to all my questions.
No, mmWave works and is used in USA and South Korea, but it is difficult to use it in high-mobility scenarios or to spatially multiplex many users, since the channels change more quickly at higher frequencies.
The channel coherence time is inversely proportional to the carrier frequency, which limits the number of data streams that can be multiplexed. Here is a blog post where Erik explains this in detail: ma-mimo.ellintech.se/2017/03/24/relative-value-of-spectrum/
It is not a strict technical constraint, but a practical constraint on how many users can share the same time-frequency resource. When K approaches M, the interference becomes too large and it is better to serve fewer at the time to maximize the total data rate. It won’t be a practical problem in 5G since we usually only have 1-2 active users at the time in current networks. Hence, if 5G Massive MIMO can handle 8 users, we have enough capacity for 5-10 years growth of data traffic.
Yes, that is likely the case, but it all depends on how the system is built (transceiver hardware) and how propagation environment look like (LOS vs. NLOS).
@@WirelessFuture so, in mmWave (LOS) there is no threat of spectral efficiency decrease. is it so? if yes kindly elaborate...and thanks for reply and giving your valuable time...actually I have many misconceptions (doubts)
@@niravpatel3961 The answer depends on the assumptions. The pathloss is the same in all bands if one consider an isotropic transmit antenna and an equal-sized receiver aperture (ma-mimo.ellintech.se/2019/10/29/is-the-pathloss-larger-at-mmwave-frequencies/). But there are other issues: Can we increase the transmit power linearly with the bandwidth to keep the SNR constant? Can we obtain equally good channel state information? What kind of arrays are we considering, etc.
@@WirelessFuture it may be possible to pump more power, but how can we get csi accurately at mmWave as coherence channel time is very short. Is there any way? because Doppler spread will also increase with higher carrier frequency. Even coherent block size will also decreases ...then how correct channel estimation will be possible... kindly reply...is there any solution for it?
According to what we know about massive MIMO and to what was said in the podcast massive MIMO performance in TDD is much better than FDD. I want to know Is not there any way to improve the performance of FDD massive MIMO and Is it true to say that the FDD should be put away?
Sure, there has been a lot of work on FDD Massive MIMO. It won’t be as good as TDD Massive MIMO, but one can still benefit benefit from having more antennas. This was discussed in the second episode of the podcast, if I remember correctly.
![Ep 22. Being Near or Far in Wireless [Wireless Future Podcast]](http://i.ytimg.com/vi/7aVCtTgyMx4/mqdefault.jpg)
![Ep 22. Being Near or Far in Wireless [Wireless Future Podcast]](/img/tr.png)
![Ep 6. Q/A on Massive MIMO [Wireless Future Podcast]](http://i.ytimg.com/vi/yDm05nor_3Y/mqdefault.jpg)
![Ep 36. 6G from an Operator Perspective [Wireless Future Podcast]](http://i.ytimg.com/vi/odqBsaSBo-4/mqdefault.jpg)
![Ep 18. Ever-Present Intelligent 6G Communications (With Magnus Frodigh) [Wireless Future Podcast]](/img/n.gif)
Its a real surprise to see Prof. Thomas Marzetta.
I am interested to understanding the gap between "the limits imposed by the nature" and our state of the art technology. I suppose understanding this will give some clues for future breakthroughs. Comments on this will be interesting.
thanks for the wonderful sessions.. really appreciate the gesture
Very inspiring, thanks Professors. Expecting more podcasts like this one in the future.
Thk-you ever so much Dr Emil Björnson. I just watched a few of your older videos & I finally got a handle on the MIMO big-picture. Of course the details still make my head explode (very messy ;). But as long as I keep-at-it, I'm sure that'll happen much less often. Of course I'll download your book too, thks-again. And I thought Sweden was only well-known for IKEA & ABBA. I wonder what else you Swedishs are great at?
A very useful podcast. thanks, Professors.
0:06:18 Channel State Information isn’t Everything; it’s the Only Thing
It was very interesting to learn about holographic MIMO. In millimeter wave imaging, the scattered field at the image plane is reconstructed by Fourier transform to obtain a snapshot of the object plane with the depth information included. I am curious how this kind of reconstruction can be performed in the context of wireless communication?
Thanks for sharing
I was really enjoying Your discussion. Let me dare ask You following question: You referred a lot to Aperture. I am seasoned Telco consultant and passionate photographer as well - Can Aperture in antennas be used in antennas in same way is in camera lenses? In photography its not fixed and static, rather Aperture is dynamic, where via adapting blades You are manipulating light thruput via lens according Your needs. Do You see any benefit doing same for antennas apertures?
There are some similarities between these concept. The aperture of an antenna array is basically its physical size and we are usually using all the antennas to maximize the signal strength. We are not changing the aperture.
But suppose we would turn off the outer antennas in the array to reduce the aperture. This will lead to wider beams and larger beam depth (if we operate in the radiative near field), at the cost of a power reduction. This is similar to what happens in a camera: you reduce the aperture to get a larger depth, but pay the price in terms of less light.
There can be a benefit of reducing the array aperture when you want the beam to cover the entire coverage area. But there are more clever ways to do the same thing using the entire array to transmit a wide beam. It is like cutting the array into many smaller pieces, operate each with a smaller aperture and then combine them to not lose any power.
Thanks you a lot for this amazing rich podcast! I have one question though!
when the discussion was about the usage of massive MIMO in Thz bands tom wasn't so positive about it.
So one can say that massive MIMO is not necessary to be used within the very high frequency bands?
I believe that Tom have spatial multiplexing of many user devices in mind when he talks about Massive MIMO. In THz bands, you will for sure need many antenna elements to get a decent link budget, but perhaps not use them for spatial multiplexing since the range will be short so the chance of having many users is fairly small.
awesome!
sir, how delay spread is inversely proportional to carrier frequency?
Strictly speaking, the delay spread isn't frequency dependent. However, when you change the carrier frequency and the array structure, the relative strength of the different paths might change and effectively reduce the delay spread. That said, I don't think it will be inversely proportional to the carrier frequency. The point that was made in the podcast is that the coherence time is inversely proportional to the carrier frequency, which will reduce the size of a coherence block, unless the delay spread also is inversely proportional (which it probably isn't)
Sir, so in mmWave technology channel estimation is a unrealizable dream? because channel coherence time decrease by 10 microsecond. kindly reply to all my questions.
No, mmWave works and is used in USA and South Korea, but it is difficult to use it in high-mobility scenarios or to spatially multiplex many users, since the channels change more quickly at higher frequencies.
sir, how spectral efficiency is inversely proportional to carrier frequency?
The channel coherence time is inversely proportional to the carrier frequency, which limits the number of data streams that can be multiplexed. Here is a blog post where Erik explains this in detail: ma-mimo.ellintech.se/2017/03/24/relative-value-of-spectrum/
please use some keywords for your videos when you upload them on youtube to reach many people. ✌✌✌✌
Why does the constraint exist for M(antennas)>>K(users)?
It is not a strict technical constraint, but a practical constraint on how many users can share the same time-frequency resource. When K approaches M, the interference becomes too large and it is better to serve fewer at the time to maximize the total data rate. It won’t be a practical problem in 5G since we usually only have 1-2 active users at the time in current networks. Hence, if 5G Massive MIMO can handle 8 users, we have enough capacity for 5-10 years growth of data traffic.
sir, so in mmWave spectrum spectral efficiency decreases?
Yes, that is likely the case, but it all depends on how the system is built (transceiver hardware) and how propagation environment look like (LOS vs. NLOS).
@@WirelessFuture so, in mmWave (LOS) there is no threat of spectral efficiency decrease. is it so? if yes kindly elaborate...and thanks for reply and giving your valuable time...actually I have many misconceptions (doubts)
@@niravpatel3961 The answer depends on the assumptions. The pathloss is the same in all bands if one consider an isotropic transmit antenna and an equal-sized receiver aperture (ma-mimo.ellintech.se/2019/10/29/is-the-pathloss-larger-at-mmwave-frequencies/). But there are other issues: Can we increase the transmit power linearly with the bandwidth to keep the SNR constant? Can we obtain equally good channel state information? What kind of arrays are we considering, etc.
@@WirelessFuture it may be possible to pump more power, but how can we get csi accurately at mmWave as coherence channel time is very short. Is there any way? because Doppler spread will also increase with higher carrier frequency. Even coherent block size will also decreases ...then how correct channel estimation will be possible... kindly reply...is there any solution for it?
@@WirelessFuture how the kind of arrays will play a role here...means is it matters here...if so kindly elaborate... thanks for reply sir...
According to what we know about massive MIMO and to what was said in the podcast massive MIMO performance in TDD is much better than FDD. I want to know Is not there any way to improve the performance of FDD massive MIMO and Is it true to say that the FDD should be put away?
Sure, there has been a lot of work on FDD Massive MIMO. It won’t be as good as TDD Massive MIMO, but one can still benefit benefit from having more antennas. This was discussed in the second episode of the podcast, if I remember correctly.
Interesting
Sir shout out please.