maki/openairinterface5g
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Commit to compare -u0 and -u2 performance in interference free environment for

Browse Source 
TM4 and verify if any additional gain/penatly is needed.
This commit is contained in:
lukashov
2016-01-05 15:27:15 +01:00
parent 73c76d1456
commit 7a8e3a421c
1 changed files with 15 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
openair1/PHY/LTE_TRANSPORT/dlsch_demodulation.c
+15 -5
View File
@@ -439,10 +439,18 @@ int rx_pdsch(PHY_VARS_UE *phy_vars_ue,
if (rx_type>rx_standard) {
// interf_unaw_shift_tm4_mcs is to avoid tails in SNR/BLER curves
avg[0] = log2_approx(avg[0]) - 13 + offset_mumimo_llr_drange[dlsch0_harq->mcs][(dlsch1_harq->Qm>>1)-1];
lte_ue_pdsch_vars[eNB_id]->log2_maxh0 = (log2_approx(avg[0])/2) +interf_unaw_shift_tm4_mcs[dlsch0_harq->mcs];//+offset_mumimo_llr_drange[dlsch0_harq->mcs][(get_Qm(dlsch1_harq->mcs)>>1)-1];
lte_ue_pdsch_vars[eNB_id]->log2_maxh1 = (log2_approx(avg[0])/2) +interf_unaw_shift_tm4_mcs[dlsch1_harq->mcs];//+offset_mumimo_llr_drange[dlsch1_harq->mcs][(get_Qm(dlsch0_harq->mcs)>>1)-1];
// !!!!! Right now in testing mode for -gS channel ONLY. We investigate if any
// additional gain/penalty needs to be introduced in comparison with I-UA receiver
// by comparing performances if run with -u2 and -u0.
// this is valid only if same mcs are used. We calibrate for mcs 4. Best shift value is 3.
// MCS-dependent LUT will be introduced.
avg[0] = (log2_approx(avg[0])/2) - 13 + interf_unaw_shift;
lte_ue_pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg[0],0);
lte_ue_pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg[0],0);
//avg[0] = log2_approx(avg[0]) - 13 + offset_mumimo_llr_drange[dlsch0_harq->mcs][(dlsch1_harq->Qm>>1)-1];
//lte_ue_pdsch_vars[eNB_id]->log2_maxh0 = (log2_approx(avg[0])/2) +interf_unaw_shift_tm4_mcs[dlsch0_harq->mcs];//+offset_mumimo_llr_drange[dlsch0_harq->mcs][(get_Qm(dlsch1_harq->mcs)>>1)-1];
//lte_ue_pdsch_vars[eNB_id]->log2_maxh1 = (log2_approx(avg[0])/2) +interf_unaw_shift_tm4_mcs[dlsch1_harq->mcs];//+offset_mumimo_llr_drange[dlsch1_harq->mcs][(get_Qm(dlsch0_harq->mcs)>>1)-1];
//printf("TM4 I-A shift layer1 = %d\n",interf_unaw_shift_tm4_mcs[dlsch0_harq->mcs]);
//printf("TM4 I-A shift layer2 = %d\n",interf_unaw_shift_tm4_mcs[dlsch1_harq->mcs] );
@@ -450,7 +458,9 @@ int rx_pdsch(PHY_VARS_UE *phy_vars_ue,
}
else
// to avoid tails in SNR/BLER curves
// to avoid tails in SNR/BLER curves. -13 is needed to make shift a positive number.
// this is valid only if same mcs are used. We calibrate for mcs 4. Best shift value is 3.
// MCS-dependent LUT will be introduced.
avg[0] = (log2_approx(avg[0])/2) - 13 + interf_unaw_shift;
lte_ue_pdsch_vars[eNB_id]->log2_maxh0 = cmax(avg[0],0);
lte_ue_pdsch_vars[eNB_id]->log2_maxh1 = cmax(avg[0],0);