Це відео не доступне.
Перепрошуємо.
Channel Estimation for Mobile Communications
Вставка
- Опубліковано 7 лис 2021
- Explains the basics of Channel Estimation for mobile communications, including time varying and frequency varying channels.
Related videos: (see iaincollings.com)
• Quick Introduction to MIMO Channel Estimation • Quick Introduction to ...
• What is an Adaptive Step Size in Parameter Estimation? • What is an Adaptive St...
• What are Flat Fading and Frequency Selective Fading? • What are Flat Fading a...
• What are Fast Fading and Slow Fading? • What are Fast Fading a...
• What is Rayleigh Fading? • What is Rayleigh Fading?
• What is Rician Fading? • What is Rician Fading?
• What is Intersymbol Interference ISI? • What is Intersymbol In...
• Sampling Signals • Sampling Signals
• How is Data Sent? An Overview of Digital Communications • How is Data Sent? An O...
Full categorised list of videos with pdf Summary Sheets: iaincollings.com
.
A very under estimated channel. I hope soon you will get the appropriate appreciation you deserve. Thank you very much.
Thanks for your nice comment. I'm so glad you like the channel.
Most university professors indulge in explaining through mathematical aspect however this is the practical explanation...Awesome.......
Glad you like the approach and found it helpful.
U are great resource for engineers
Thanks for your nice comment. I'm glad people are finding the videos useful.
Excellent and lucid explanation. Would like to know how decision directed estimation (data basically) in good channels helps in the channel interpolation .. Again thanks a lot
Well, if you detect/decode the data correctly, then you can use it in exactly the same way you use pilot symbols, to estimate the channel. The data can act as extra pilot symbols. If you are able to detect/decode all of the data correctly, then you don't even need pilot symbols at all. In OFDM, some sub-channels could use low order QAM sending data at a low rate with reliable data detection/decoding (the detected/decoded data could be used to add to the pilot symbols to better estimate the channel), while other sub-channels could use higher order QAM (to get the higher overall data rates).
@@iain_explains Thanks a lot for the reply. My understanding is that pilot is a known signal and can be used for channel estimation whereas data is random (i.e. random constellation point) . What I am unable to comprehend is how data (even though low modulation order like bpsk or qpsk) can be used for channel estimation. In lte for e.g. unfortunately all carriers (RE) allocated to a user will have same modulation order unlike say ADSL where each carrier can have different bits allocated (from 2 bits to 15 bits). Can you please elaborate on this? Or is this something like decision feedback Equalizer concept where previous decoded data symbols are used to estimate current symbol .. Am I missing something ?
@@amitpalkar934 I have similar confusion. Please post if you get an answer?
I'm a computer engineering student preparing a seminar on CGAN models for channel estimation.
Could you explain why channel estimation is important?
What values are typically found in the channel matrix (like h12, h11, h22, h21)?
Also, is channel estimation used in everyday phone calls to obtain the channel matrix for equalization?
I would appreciate your insights!
@iain_explains Sir, I've watched the video you mentioned in the link, but there's no explanation about the channel coefficients inside the channel matrix.
It would be easier for me to understand if you could reply with answers to the 3 questions I asked as a reply comment (explanation is only required on a surface level and in simple words)
What book do you recommend to see topics of MIMO systems, OFDM and mobile communications? Great video.
Here are a couple of suggestions:
• B. Clerckx and C. Oestges, "MIMO Wireless Networks: Channels, Techniques and Standards for Multi-Antenna, Multi-User and Multi-Cell Systems"
• Y.S. Cho, J. Kim, W.Y. Yang and C.G. Kang, “MIMO-OFDM Wireless Communications with Matlab”
@@iain_explains Thank you!!!
sorry for this dumb comment - in the GSM case, once you get the channel estimate, do you have to go to frequency domain to cancel out the channel effect? I am thinking (sorry for the simplistic and non-engineering view) it will be just a matrix division there rather than the process of deconvolution? please comment (i get it in the OFDM case but not sure in GSM ?)
It's been a while since I was thinking about 2G GSM systems. But basically, GSM is what we would now call "narrow band", and it used "single carrier" modulation. So the channel was effectively "flat" in the frequency domain. The main thing to worry about was the time variations in the channel. These videos might shed more light: "Mobile Standards Evolution: FDMA, TDMA, CDMA, OFDMA" ua-cam.com/video/bm53RpK-S2k/v-deo.html and "What are Flat Fading and Frequency Selective Fading?" ua-cam.com/video/KiKPFT4rtHg/v-deo.html
Thank you for the informative video! I am currently studying channel estimation for 5G for my Thesis. I would like to ask a question, and I apologize in advance for my simple question. I searched about the signal/energy value used for simulating pilot symbols in matlab. From what I currently know, information bits in matlab is simulated by randomizing bits 0 and 1. Pilot symbols however does not used bits (as far as I know) and the value is known at both the transmitter and the receiver. What is the normally used signal/energy values used for pilot symbols for channel estimation? Thank you in advance!
The power used for the training signal depends on the particular specification in the relevant standard. There is no "theoretical optimal" power.
hi lian,
suppose if we have coherence time Tc for a time varying channel than for that particular duration weather we are sending our data and pilot bits or we are only sending pilot bits for the estimation of channel.
Over that duration you can assume that the channel doesn't change too much (more or less - like I say, it's a vague concept). So within that time you need to send some training symbols to help with channel estimation and send data symbols that will be detected and decoded using that channel estimate. After that time, the channel will likely have changed, so you'll need to send some more training symbols and re-estimate the channel, and so on.
I found this video hard to follow relative to your other videos. Is there a sequence of videos you would recommend for building up an understanding of this topic?
It's hard to answer this question, without knowing what you've watched/understood already. Perhaps check out some of the Playlists on the channel. I've tried to sequence the videos in an order that builds on ideas. Also, check out my webpage, were each topic has videos listed in order of developing ideas: iaincollings.com
Can you make a video on channel reciprocity in wireless time-division duplexing systems for channel estimation?
Thanks for the suggestion. I've added it to my "to do" list.
Where did you get the least square estimate of h equation from? Is this covered in one of your other video lectures?
Thank you
I don't think I've derived it explicitly anywhere, but these two videos will hopefully go some way to providing the answer: "What is Least Squares Estimation?" ua-cam.com/video/BZ9VlmmuotM/v-deo.html and "How are Different Equalization Methods Related? (DFE, ZF, MMSE, Viterbi, OFDM)" ua-cam.com/video/hYNwTTWrp48/v-deo.html
Very useful as usual. Thank you. On the subject of channel estimation, a video on SU-MIMO and MU-MIMO channel estimations to help beamforming might also be useful. I get how MIMO channel estimation can be achieved when a different pilot is used for each antenna/user/port (5G terminology), basically this would be the same as your explanation in this video. However, let's say we have a 4x4 MIMO and only one pilot sequence is used, eg Gold Sequence. How do we get the 16 channel parameters out of the single sequence??
PS: Assuming channel reciprocity, MISO and SIMO is also easy to get with a single pilot sequence.
Thanks for the suggestion. I'll add it to my "to do" list.
What can I say! Just u r a creative man .
How can I add Doppler shift to the channel to avoid fast fading
Sorry, I'm not sure what you mean by "adding" Doppler shift. Doppler shift occurs when there is relative movement between the transmitter and receiver, or by the scatterers. You don't "add" it. This video may help: "What are Doppler Shift, Doppler Spread, and Doppler Spectrum?" ua-cam.com/video/LLr3-kotbz4/v-deo.html
What about channel estimation by using IRS and deep learning
Sorry, I'm not sure what you're asking exactly. Yes, you can throw "deep learning" at pretty much any identification/estimation problem and see what it gives you.
Hello sir , if there is no relative motion and no Dopplers than also why we need to periodically transmit our pilot bits for estimation of channel, I mean to say channel response is not changing with time so we should only estimate our channel once and then we can detect our signal at the receiver end
How do you know there won't be any relative motion? If you're designing a system for a mobile terminal, then you need to design your transmission protocol to allow for motion.
@@iain_explains ok sir i got your point but consider a case when you are talking to a person over mobile phone for an hour remaining stationary (sitting on your working desk) with no motion at all then the channel is stationary so in this case also there is estimation of channel required again and again? or one time estimation would be enough?
Well the network/basestation doesn't know that you intend to stay perfectly still at your desk. It needs to be prepared for you to start moving around at any moment. If it doesn't have continual 'channel sounding' then it won't be able to work out if you start moving, and then respond. It would result in your connection going into a high error state.
@@iain_explains thanks professor it cleared my most of my confusion
please make a video on the log-likelihood ratio for qpsk...with a numerical example.
Thanks for the suggestion. But I'm not sure which log-likelihood ratio you are talking about, sorry.
@@iain_explains sir, suppose we have QPSK symbols mapped to {+-1,+-j}, bits " b0b1" passed through a scalar fading channel :y=hx+n, where n is Gaussian with mean=0 and varience (sigma)^2.
Then the value of
LLR(b0)= log[pr(y/b0=0)/pr(y/b0=1)]
will be..
Sir pls provide insight into it.
Well, if you just want the LLR for b0, then you'd have to average over the possible values of b1. You can't have pr(y|b0) without also saying what you're going to assume about b1 - either by conditioning on its value, or averaging over its pdf.
I am after the channel estimation for low power devices (wireless sensor network)
How realistic it is to assume channel reciprocity in TDD?
Great question. It depends on the electronics and antennas that are being used at each end. If their effect can be "calibrated out" then TDD should work pretty well. It's being used in some standards already.
I've been hearing a lot of buzzwords around WiFi CSI lately. What exactly is this? I couldn't find much stuff on internet, all sources seem to be reiterating same points
Thanks for the suggestion. I've added it to my "to do" list.
"Might" be similar to CSI-RS in 5G-NR which are REFERENCE-SIGNALS transmitted in DL helping the user equipment in measurement of the DL channel which can be reported back to the base station for tuning and adaptation techniques which will improve the efficiency of data transfer. It is used for multiple other features in 5g-nr.