实际系统中，物理层发射和接收的是调制在载波上的射频信号，但是由于运算量和复杂度的原因，在仿真系统中物理层不可能采用和实际系统一样的收发方式。经过物理层抽象，发射和接收的是未经过基带处理和射频调制的MAC层报文，以及表征信号特征的若干参量，包括发射、接收信号功率、天线增益、热噪声功率、干扰功率等，物理层利用这些参量可以得出本地接收信号的SINR。而物理层特性，包括基带信号处理增益（信源编解码、信道编解码、扩频、调制等）、射频增益、频偏、干扰、衰落等被统一抽象为SINR-BER特性。
In OFDM system, the link-to-system performance mapping is required to translate the effective channel quality index, e.g. SINR or RBIR of each sub-channel, into the performance metric, e.g. block error rate(BLER). Note that the received signal quality by each sub-carrier in the same sub-channel is usually different.
There are several well-known popular mapping functions proposed in literatures, such as:
- linear average over the frequency selective fading channel
- exponential effective SINR mapping (EESM)
- mutual information based link quality model