[chai]

Hi

When try to run FBMC modulation technique, this error pops out “Frame durations of cells do not match!”

It happens for both genericscenario and numerology, can anyone help me with this?

[Stefan Pratschner]

Dear Chai,

as described in the documentation, our simulator works on a frame basis. Therefore, frames of all cells need to have an at least approximately equal frame duration. This becomes critical when different numerologies or waveforms are used in different cells.

What parameters are you using for the two cells? Did you change anything other than the waveform in the two mentioned scenarios?

best
Stefan

[chai]

Dear Mr Stefan,

i used the parameter in the numerology and the parameter are as below:
scStr.modulation.subcarrierSpacing = [15e3, 60e3]; % select a four times bigger subcarrier spacing in the second cell
scStr.modulation.numerOfSubcarriers = [208, 52]; % to have the same bandwidth, the total number of subcarriers must be one fourth on the second cell
scStr.modulation.nSymbolsTotal = [15, 60]; % to have the same frame duration, the total number of symbols must be four times higher in the second cell
scStr.modulation.nGuardSymbols = [1, 4];

the same case happens when i try using the ‘genericscenario’:
scStr.modulation.numerOfSubcarriers = [72, 72]; % per BS; number of used subcarriers
scStr.modulation.subcarrierSpacing = [15e3, 15e3]; % per BS; per base station in Hz
scStr.modulation.nSymbolsTotal = [15, 15]; % per BS; total number of time-symbols per frame, the frame duration will be nSymbolsTotal/subcarrierSpacing
scStr.modulation.nGuardSymbols = [1, 1];

Chai

[chai]

Dear Mr Stefan,

I used the same parameter for OFDM, f-OFDM and UFMC. In this case why is FBMC not working? (How does the frame duration does not match with each other but it match with the other 3 modulation)

Please reply soon. Thank you.

Chai

[Stefan Pratschner]

Dear Chai,

the problem related to the frame duration is, that different waveforms apply different filters in time domain. However, the whole simulator processes data on a frame-basis. Therefore, if different waveforms are applied in different cells, they still have to be similar across cells such that the frame processing is still possible.

I am sorry for my late reply. I will look into this issue. Are you using a single waveform, that is, FBCM, in all cells?

best
Stefan

[Stefan Pratschner]

Dear Chai,

thanks for your comment. Unfortunately there was a bug concerning the use of FBMC. I fixed this bug and re-uploaded the files. Please consider to re-download the simulator files. There were changes in the User.m as well as in the simulationParameters.m.

However, in the case of FBMC, only real valued symbols are transmitted. To obtain the same time duration as in a cell with OFDM modulation, the number of symbols in the FBMC cell has to be twice as large compared to the OFDM one. So, to obtain LTE like settings, you should use 15 symbols in the OFDM cell while you set 30 symbols in the FBMC cell. Further, as FBMC has no guard symbol (or CP), you should set the number of guard symbols for the FBMC cell to zero.

Hope that solves your problem.

best
Stefan

[chai]

Dear Mr. Stefan,

After i redownload the simulator and used all the setting you mention, unfortunately it still doesn’t work. I still cannot run a simulation comparing OFDM with FBMC.

Below are the parameter i used:
“scStr.modulation.subcarrierSpacing = [15e3, 30e3]; % select a four times bigger subcarrier spacing in the second cell
scStr.modulation.numerOfSubcarriers = [104, 52]; % to have the same bandwidth, the total number of subcarriers must be one fourth on the second cell
scStr.modulation.nSymbolsTotal = [15, 30]; % to have the same frame duration, the total number of symbols must be four times higher in the second cell
scStr.modulation.nGuardSymbols = [1, 0]; ”

it comes with an error:”Error using Parameters.SimulationParameters/checkParameters (line 173)
Invalid ratio of subcarrier spacing, number of subcarriers, number of total symbols and number of guard symbols across
cells.”

and the same case where i use 1 symbols and 2 symbols

Chai

[Stefan Pratschner]

Dear Chai,

as I already explained, the number of FBMC symbols has to be doubled compared to any other waveform, as it uses only real valued symbols. Further, as you employ a subcarrier spacing for cell 2 that is twice as large as cell 1, you again have to double the number of symbols to match the frame duration. Therefore, I suggest the following parameters

scStr.modulation.subcarrierSpacing = [15e3, 30e3]; % select a four times bigger subcarrier spacing in the second cell
scStr.modulation.numerOfSubcarriers = [104, 52]; % to have the same bandwidth, the total number of subcarriers must be one fourth on the second cell
scStr.modulation.nSymbolsTotal = [15, 60]; % to have the same frame duration, the total number of symbols must be four times higher in the second cell
scStr.modulation.nGuardSymbols = [1, 0];

I just tried these settings on the current simulator version and it worked. Let me know if you encounter further problems.

best
Stefan

[chai]

Dear Mr. Stefan,

I follow exactly the same parameter u suggested and redownload the simulator, i still get an error of

“Error using zeros
NaN and Inf not allowed.

Error in Channel.FastFading (line 117)
PowerDelayProfileTemp = zeros(length(IndexDelays),max(IndexDelays));

Error in Parameters.SimulationParameters/checkParameters (line 212)
Channel.FastFading( obj.modulation.samplingRate, … % sampling rate

Error in Parameters.SimulationParameters (line 27)
obj.checkParameters();

Error in main (line 19)
simParams = Parameters.SimulationParameters( simulationScenario );”

what should i do?

Chai

[Stefan Pratschner]

Dear Chai,

I just tried again; using the currently available simulator with the following changes in the generic scenario:

scStr.modulation.waveform                   = { 'OFDM', ...             % waveform for each BS:
                                                'FBMC' }; 

scStr.modulation.subcarrierSpacing = [15e3, 30e3]; % select a four times bigger subcarrier spacing in the second cell
scStr.modulation.numerOfSubcarriers = [104, 52]; % to have the same bandwidth, the total number of subcarriers must be one fourth on the second cell
scStr.modulation.nSymbolsTotal = [15, 60]; % to have the same frame duration, the total number of symbols must be four times higher in the second cell
scStr.modulation.nGuardSymbols = [1, 0];

scStr.modulation.samplingRate               = 30e3*104*2;

scStr.schedule.fixedSchedule{1}             = [ 'UE1:104' ];     % schedule for BS1
scStr.schedule.fixedSchedule{2}             = [ 'UE2:52' ];     % schedule for BS2

works perfectly fine. It seems the error occurs since you are using automatic sampling rate calculation.

Thank you for your comment, we will fix this issue. For now I advise to use the largest subcarrier spacing times the maximum number of subcarriers times two as sampling rate.

best
Stefan

[chai]

Thank you so much Mr. Stefan, i works.

last question from me, in the result of BER, there is this UE1, UE2, coded and uncoded, can i know what does the coded and uncoded stands for?

Chai

[Stefan Pratschner]

Dear Chai,

uncoded means that there is no channel coding involved, while coded means there is. So for the coded BER, the data bits before the channel encoder at the transmitter are compared to the decoded bits after the decoder at the receiver side. For the uncoded BER, the encoded bits at the transmitter after the channel coder are compared to the received bits before the channel decoder at the receiver.

best
Stefan

[Author]

Posts