SINR measurements between connected and unconnected entities

[Diego Bezerra]

Hi, I am new to the ns-3 and need help. I have a scenario with 1 UE and few eNBs. I would like to read the SINR measurements (from a eNB to which the UE is connected and to one where it is not).

I know that through the "/NodeList/*/DeviceList/*/$ns3::LteUeNetDevice/ComponentCarrierMapUe/*/LteUePhy/ReportCurrentCellRsrpSinr" route I can get the values (in mW) for the SINR considering the currently connected eNB. However, this measure is not presented for the other eNBs (not connected).

I know I can still calculate the SINR from the obtained RSRQ, since this measure is presented for both cases. I did tests using the measurements obtained (in dB) in "/NodeList/*/DeviceList/*/$ns3::LteUeNetDevice/ComponentCarrierMapUe/*/LteUePhy/ReportUeMeasurements". However, the equations that can be used (https://www.cablefree.net/wirelesstechnology/4glte/lte-rsrq-sinr/) have limitations when the RSRQ values are high.

Could you please tell me how to do it.

Thanks, Diego.

[Diego Bezerra]

Hi everybody,

I've been investigating the reported issue.

In the LteUePhy class, the NS-3 presents the ReceivePss function (receive Primary Synchronization Signal) where the id of each cell in the scenario (cellId) and the Power spectral density (PSD) (W/Hz) (p) are passed as arguments. The power value measured between the UE and the eNB is used to calculate the RSRP. In this process, the PSD is converted to linear power (W).

Knowing that the SINR is given by S/(I+N), for LTE measurement, S is the average received power of the reference signal (known as RSRP). For the denominator, we have a sum between interference and noise power. Based on this, it was possible to identify that in the GenerateCqiRsrpRsrq function the interference and noise measures are computed to calculate the RSRQ. This measurement is stored in the interfPlusNoisePowerTxW variable and based on m_rsInterferencePower.

Once we have a way to measure S, which gives rise to the RSRP, and measure the interference and noise received, we made the appropriate conversions to W, applied these measurements in the SINR formula, and converted to dB (in ReceivePss function).

A scenario was modeled considering the user mobility and the existence of three eNBs (blue, orange, and green). During the simulation, the behavior of the SINR calculated as described above is shown in the Figure below. As a simple test, the user travels at 200 m/s (I know it's unrealistic. The goal was to speed up the simulation). According to the table that indicates LTE signal strength values, a horizontal line defines a threshold at 20 dB, indicating when the SINR is excellent. When connecting with the first eNB (blue), the SINR calculated for the other eNBs is low. As it approaches, this value tends to increase until the handover trigged.

When comparing the results obtained with the SINR delivered by the NS-3 in the Figure below (only between the eNB to which the UE is connected), we can note that the behavior is very similar, although the values obtained are different (min: 6.31 dB, max: 108 dB obtained from formula vs min: 51.53 dB, max: 144.46 dB reported by NS-3).

Having said that, how do you rate it? Has anyone had to deal with a similar problem and got a different solution?

Thanks,

Diego