Questions
Welcome to IWSLS² 2018 › Forums › IWSLS² 2016 › Paper Discussion Corner › Filter Bank Multi Carrier Based MIMO System for 5G Wireless Communication › Questions
This topic contains 1 reply, has 2 voices, and was last updated by Presenter 2 years, 8 months ago.

AuthorPosts

July 1, 2016 at 2:21 pm #298
Thanks your for your presentation. I have got a few questions:
– I suppose that is a typo, but why are some symbols “s” in Fig. 1 with tilde and others not?
– When you talk about offset QAM, you seem to claim that you achieve twice the rate of OFDM with the same bandwidth or did I misunderstand that? According to my knowledge of OQAMFBMC, it achieves the same data rate by reducing the spacing in frequency by a factor of two but transmitting only purely real/imaginary symbols.
– I did not understand your explanation of frequency selectivity: do you mean that the Rayleigh fading channel realization is independent from one subcarrier to the next?
– In Eq. (9), what is the subscript n?
– Your interpretation of OQAM in Fig. 6 is somewhat surprising: OQAMFBMC normally means that you transmit on one FBMCsubcarrier the real part of a QAMsymbol and on the consecutive subcarrier you transmit the imaginary part of the same QAMsymbol. You, however, change the entire symbol constellation. Why is that?
– In OQAMFBMC, intercarrierinterference (ICI) is purely real/imaginary (depending on the subcarrier index). Due to this, ICI in SISOOQAMFBMC can be avoided by dividing by the channel coefficient and taking the real/imaginary part of the channel output. It is well known that this approach, however, fails in MIMO since it cannot cancel the real/imaginary interference from multiple transmit antennas as the channel coefficients from these antennas are different. In your approach, this problem does not seem to appear since you don’t consider the ICI channel at all; yet, in contrast to OFDM, ICI in the complex domain is always there in FBMC. Why does it not pop up in your system model?Thanks.
 This topic was modified 2 years, 8 months ago by Participant.
July 10, 2016 at 12:52 pm #321First of all I want to thank you for taking such keen interest into my presentation being this my first ever paper, it means a lot for a budding researcher like me.
I will try to answer your questions in the order you posted them.
– I suppose that is a typo, but why are some symbols “s” in Fig. 1 with tilde and others not?
Yes this is a typing mistake and it happened in the formatting. and I apologize for that. Please consider all the “S” in figure 1 as same, the normal one and the ones with the tilde they are all the same.– I did not understand your explanation of frequency selectivity: do you mean that the Rayleigh fading channel realization is independent from one subcarrier to the next?
– In Eq. (9), what is the subscript n?
About the Frequency selectivity question, In our experiment the data is grouped into blocks and then 1 block is transmitted in each time slot. so for every block of data transmitted, the channel parameters change, so, yes each block of data is effected independently from the by the channel.
The ‘n’ is the symbol index for the pair of symbol broken up into real and imaginary component.
– When you talk about offset QAM, you seem to claim that you achieve twice the rate of OFDM with the same bandwidth or did I misunderstand that? and the question about figure 6
First of all its not a comparison between OFDM and FBMC , its a comparison betweem QAM and OQAM. And the term used here is data symbol density not data rate. The question about figure 6 also gets answered here as you can see the output of two modulators in it. you can see how QAM symbols are mapped with gaps like on 1,3,5,7 as compared to that OQAM (PAM Based) symbols are mapped on positions 1,2,3,4,5….and so on and there are symbols overlapping in the OQAM mapping but 2D constellation is not capable of depicting that.
so we have 4 axis here the positive and negative real and imaginary axis so density is increased four times. but with PAM having the ability to carry half the information content only as compared to QAM. The data symbol density is said to be increased twice.(my claim here is referenced with [11], page 102103 IEEE signal processing magazine MAY 2011)– In OQAMFBMC, intercarrierinterference (ICI)
I agree with your point about FBMC performing badly with MIMO. But it shows good results with Multitone transmission methods such as FMT, CMT and SMT.
we used SMT in our study, so we have real and imaginary symbols alternating on the subcarriers making them orthogonal, along with that data encoding technique used is transmitting data in the form of blocks. and the equalization techniques used here work of eliminating the component of inference signal picks up from the other symbols, that also improves the performance. along with that each revived antenna wave form is demodulated separately. And the either real or imaginary part of the data is taken for calculating BER only.I hope i was able to satisfy your question, if you have anything else you want to know please feel free to post here or contact me on my email address for more detailed discussion.
best regards
ATIF BASHEER
INSTITUTE OF SPACE TECHNOLOGY ISLAMABAD PAKISTAN
atif_basheer 1990@hotmail.com 
AuthorPosts
You must be logged in to reply to this topic.