Basic Channel Capacity Behaviors [Video 4]

In this video, Associate professor Emil Björnson explains how the channel capacity depends on the signal power and bandwidth. This video is a part of the series Introduction to Multiple Antenna Communications: • Introduction to Multip...

[@samant6812]

sir your lectures are very much useful.

[@thinyuaung9333]

Thank you so much...very nice explanation.

[@wafaelhajhmida3863]

Dear Mr. Bjornson, You have an incredible way to convey your knowledge :) Could you plz make for us a video to explain why the data rate is related to the frequency bandwidth using Fourier transform ? Thx a lot

[@WirelessFuture]

You don’t need to involve the Fourier transform. Just apply the sampling theorem, which says that a signal is described by a number of samples proportional to the bandwidth. These samples are used to carry information, for example, by designing the signals using pulse-amplitude modulation. It is briefly mentioned in Video 3 and around slide 22 in kzbin.info/www/bejne/iYOUYXmbntp0fNU

[@wafaelhajhmida3863]

@@WirelessFuture thx a lot

[@khalidshaikh1098]

Hello Thankyou for the detailed explaination - I have one question If I have a Wi-Fi setup which provides me the following information at the receiver side - RSSI (signal strength at the receiver) and the Noise floor in dB and the channel frequency response is there any way i can evaluate the capacity of the channel with the above values as I have no acess to baseband signal to get the energy of each bits.?

[@WirelessFuture]

The easiest thing would be to compute the SNR as the RSSI divided by the Noise power (be careful with the dB-scales). You can then estimate the capacity as "Bandwidth*log2(1+SNR)".

[@khalidshaikh1098]

@@WirelessFuture Thankyou for the explaination , just a follow up question : Considering a 2x2 MIMO channel , I have the channel state information of received signal and RSSI and Noise Floor values of each Tx-Rx path, which formulae can I used to estimate the capacity of Wi-fi system ? considering rest other OFDM parameters (MCS, Bandwidth...etc ) is known to me. It would be reallly help- ful for my Thesis work. Thankyou in advance :)

[@abinetendale2455]

Nice video

[@juliocarossella6026]

00:48 is possible to have a non-symmetric two-sided spectrum??

[@WirelessFuture]

Yes, but then the time-domain signal will be complex valued, so such a signal doesn’t exist in practice. However, the non-symmetric complex baseband signal exist and is usually represented as two different signals. (The “in-phase” and “quadrature” signals)

[@juliocarossella6026]

@@WirelessFuture ... but the graph is representing a baseband 2-sided spectrum.... is that possible?

[@XTQ1179]

The Shannon limit says there is an upper bound for a reliable communication. Then what is the meaning of "reliable communication"? In other words, if one exceeds the limit, how much deterioration will be brought about?

[@WirelessFuture]

Suppose we transmit a block of N modulation symbols (samples). If their information content is below the Shannon limit, there is a way to encode the information (select the symbols) so that the probability of incorrect reception of the block goes to zero as N increases towards infinity. If we instead exceed the limit, then as N increases, the communication will always fail (with probability approaching 1). So it is really becoming binary. Below the limit: Success. Above the limit: failure.

[@XTQ1179]

@@WirelessFuture Thank you very much! Great explanation!

[@kozhenidres314]

♥️😍 another piece of cake but the equation @3:07 should not be in (Symbol/second) instead (bits/s) because i think it's going to be bits/s after modulating the symbols isn't it ?

[@WirelessFuture]

If you pause the video at 2:34, you will see expressions in symbol/s and in bit/symbol. If we multiply them together it becomes bit/s. This what is presented at 3:07.

[@abdulazizalakoub1087]

Thank you very much B (Symbol/sec) and B (Bandwidth) , I have a confused in this abbreviations

[@WirelessFuture]

B is the bandwidth in Hz, but due to the sampling theorem it also equals the number of complex symbols per second. (The sampling theorem says that a signal with bandwidth B is uniquely described by 2B real samples per second or B complex samples.

[@lonleyangel95]

how log2 (1+z)= log2(e)!!!

[@WirelessFuture]

log2(1+z)≈log2(e)*z when z is close to zero.

[@lonleyangel95]

@@WirelessFuture okai thank you

[@donpkchannel7203]

Very bad. You talk without thinking about us understanding. Your just reading of some text. Try to be better.

[@WirelessFuture]

This is a brief overview of the concept. If you want a more in-depth description, we recommend the following video: kzbin.info/www/bejne/jIa9hJlog9GHr68 In fact, there is an entire lecture series: kzbin.info/aero/PLTv48TzNRhaJ66mW48MI_HBBawupV_ZR_