Channel estimation is a key aspect for the performance of the physical layer in all MIMO-OFDM systems. In order to obtain realistic performance assessments from the LTE link level simulator, channel estimation errors must be taken into account. Implementing a detailed channel estimation algorithm may lead to long simulation time, in particular for wideband MIMO transmission modes where many channels have to be estimated, and for procedures like Wiener filtering which require matrix multiplications. In this paper we propose to simulate channel estimation errors by a Gaussian additive noise error model. The noise variance of the model is calibrated as a function of the received SNR for different channel models and Doppler spreads. The model is validated by comparing the BLER curves obtained with the model with the curves obtained when using real channel estimation.

Peer Reviewed