 This topic has 5 replies, 2 voices, and was last updated 4 years, 5 months ago by Stefan Pratschner.

AuthorPosts

7. November 2016 at 2:12 #6539SeerwanParticipant
Dear Stefan,
Hope you are fine.
Please could you help in these questions?Q1/ Is the RRH will be included in the next version of uplink simulator and approximately what time of next version of uplink?
Q2/ Is the content of channel matrix H represents channel state information (CSI)? if yes is the CSI involves (CQI, RI, PMI)? And where in the simulator the matrix H convert to these (CQI, RI, PMI)?
Q3/ Is the communication link between the eNodeB and the UE is considered perfect without delay? and adding delay to the link (from load paramters), is that delay for the link from user to base station ?
Q4/ In the current LTEA, which estimator is used? Is it MMSE?
Q5/ Is there relation between the number of pilot symbols and the dimensions of channel matrix H? or the number is fixed number (72) for any dimension of H?
Q6/ How the throughput per user is calculated and could you show me where the equation that is used to find throughput in the simulator? (is the equation available in one of your papers) And is the throughput per cell just the sum of the throughput of all users?Many thanks
Seerwan
8. November 2016 at 11:54 #6616Stefan PratschnerKeymasterHi,
Q1. Unfortunately we do not plan to include RRH in the uplink link level simulator.
Q2. The channel matrix H consists of channel coefficients for all subcarriers and OFDM symbol times from all transmit to all receive antennas in the case of MIMO transmissions. The feedback values, namely CQI, RI and PMI are then calculated from this (estimated) channel matrix in the scheduler (lteScheduler.m).
Q3. The delay value of LTE_params.downlink_delay says after how many TTIs the calculated feedback values are exploited at the eNodeB. This emulates the delay in signalling of the CQI (and PMI,RI) value from the eNodeB to the UE.
Q4. In case you refer to the channel estimation schemes, there are various available including MMSE. Please find the corresponding description of LTE_params.BS_config.channel_estimation_method in the documentation.
Q5. In LTEA uplink there is one pilot symbol per subcarrier and slot allocated. In case the transmission bandwidth increases (and the number of subcarriers increases) also the number of allocated pilots increases, according to the 3GPP LTEA standard.
Q6. Yes the cell throughput is the sum of all user throughputs in the specific cell. It is calculated in the process_TTI_results function in simulationResultsUL.mgreetings
Stefan9. November 2016 at 20:41 #6628SeerwanParticipantDear Stefan,
Many thanks for your valuable comments. Stefan Just regarding the Q3 above in reality what is the main source of this delay (in case there is a delay in the system), is it from the complexity of the channel estimator? or from the overhead in computations or from the communication link delay between the UE and eNodeB? or from the large number of pilot samples or from the scheduler? or there is anything else? Thank you very much.
Best regards
Seerwan
11. November 2016 at 7:41 #6631Stefan PratschnerKeymasterHi,
well naturally the pilots are sent to the eNodeB in a certain TTI, even if the signal processing (channel estimation, feedback parameter calculation) is performed infinitely fast, the feedback parameters are exploited at earliest in the next TTI, meaning due to the feedback and signalling from the UE to the eNodeB and back there is inherent delay of at least 1 TTI (one can also argue that this inherent delay is even 2 TTIs).
However, the signal processing at the eNodeB, mainly channel estimation, scheduling and link adaptation, takes some time and make the major part of the delay. For LTEA 3GPP mentions a delay of 5 TTI for the feedback parameters. This is even a quite optimistic assumption in my opinion.
greetings
Stefan13. November 2016 at 23:13 #6643SeerwanParticipantDear Stefan,
I Highly appreciate your support and help.– Regarding the Q2 above (although your answer is very perfect but please pardon me for my further questions on these points). Just for more clarification, you said that “The feedback values, namely CQI, RI and PMI are then calculated from this (estimated) channel matrix in the scheduler”, this is happen in the base station however, to best of my knowledge, In order to support dynamic adaptation, these three reports (CQI, PMI and RI) should be generated by the mobile terminal UE and then feeding them back to the base stations through the (channel matrix H) to aid with scheduling and link adaptation. I thing just instead of CQI the UE calculate the SINR, so the channel matrix that has been sent by the UE to the base station already include these information. I wonder if I am right or wrong?
– Regarding Q6 above,
The cell throughput = SUM ( all UEs throughput).
UE throughput (i) = B * log(1+SINR(i))
where,B is the bandwidth and the value of SINR is given in the simulator as vector of values (SNR_vec = linspace(0,30,8).
My question is the above equation of UE throughput is same the equation that is used in the simulator to find the UE throughput? and in which line of code is calculated? because I have difficulty to find it in simulationResultsUL.m . Many thanks
Best regards
Seerwan
21. November 2016 at 9:45 #6893Stefan PratschnerKeymasterHi,
the throughput in the LTEA uplink link level simulator is not calculated via SINR. In our link level simulator random data bits are generated, encoded and modulated according to the LTEA standard. At the receiver all successfully transmitted frames are taken into account for the throughput calculation. This means we actually count how many bits are transmitted in 1ms, which is generally less than the capacity equation you proposed.
greetings
Stefan 
AuthorPosts
 You must be logged in to reply to this topic.