System Level Modeling and Optimization of the LTE Downlink
System Level Modeling and Optimization of the LTE Downlink
System Level Modeling and Optimization of the LTE Downlink
- No tags were found...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
3. Physical Layer <strong>Modeling</strong> <strong>and</strong> <strong>LTE</strong> <strong>System</strong> <strong>Level</strong> Simulationa decimation factor <strong>of</strong> six, which results in 2 · N RB SINR values for <strong>the</strong> whole <strong>LTE</strong>system b<strong>and</strong>width shown in Table 3.1. Additionally, due to <strong>the</strong> averaging nature <strong>of</strong><strong>the</strong> rank <strong>and</strong> precoding feedback algorithms [51, 54], at least two subcarrier SINRvalues per RB are desirable.Table 3.1.: Number <strong>of</strong> calculated SINR samples for <strong>the</strong> different <strong>LTE</strong> b<strong>and</strong>widths.Channel b<strong>and</strong>width (B channel ) [MHz] 1.4 3 5 10 15 20Number <strong>of</strong> SINR samples (2 · N RB ) 12 30 50 100 150 2003.1.1.1. Post-equalization SINR <strong>and</strong> Trace GenerationAs aforementioned, <strong>the</strong> complexity <strong>of</strong> <strong>the</strong> link quality model can be reduced byconsidering only a subset <strong>of</strong> <strong>the</strong> total post-equalization SINRs. For <strong>the</strong> calculation <strong>of</strong><strong>the</strong> SINR itself, a simple linear receiver, <strong>the</strong> Zero Forcing (ZF) receiver, is considered.Since system level simulations are to be used where Multi-User (MU) diversity is alsoexpected to be exploited (for single-link simulations one would ra<strong>the</strong>r employ linklevel simulations), it can be argued that for a large number <strong>of</strong> users, <strong>the</strong> ZF receiverapproaches <strong>the</strong> average performance <strong>of</strong> <strong>the</strong> optimal receiver, since MU diversityeffect can compensate for poorly conditioned channel matrices [77].Assuming perfect channel knowledge, where <strong>the</strong> accent mark [ˆ] denotes a receiverestimate, <strong>the</strong> estimated received symbol vector ˆx can be expressed as()N∑intˆx = Gy = G H 0 x 0 + n + H i x i , (3.5)where G is <strong>the</strong> receive filter, which for <strong>the</strong> ZF receiver is calculated as <strong>the</strong> pseudoinverse<strong>of</strong> H <strong>and</strong> expressed asi=1H + = ( H H H ) -1H H , (3.6)where H H denotes <strong>the</strong> Hermitian transpose <strong>of</strong> H.For <strong>the</strong> cases where a spatial multiplexing (OLSM or CLSM) or TxD (based on <strong>the</strong>Alamouti STBC [41]) mode is employed, H denotes <strong>the</strong> effective channel matrix.For <strong>the</strong> SM cases, it can be calculated as <strong>the</strong> combination <strong>of</strong> <strong>the</strong> actual channelmatrix <strong>and</strong> <strong>the</strong> linear precoder W that maps <strong>the</strong> ν transmitted symbols 2 to <strong>the</strong>N TX transmit antennas. For TxD, <strong>the</strong> precoding-equivalent operation is shown for2 In <strong>the</strong> <strong>LTE</strong> st<strong>and</strong>ard, <strong>the</strong> number <strong>of</strong> simultaneously transmitted symbols is referred to as <strong>the</strong>number <strong>of</strong> layers, <strong>and</strong> is denoted as ν (see Chapter 2 <strong>and</strong> Section 2.2.2).29