OFDM Noise Power Estimation Using Time-domain Channel Response

Abstract

In this thesis, a noise power estimation algorithm is proposed from an orthogonal frequency division multiplexing (OFDM) system. In many OFDM systems, pilots are periodically transmitted for channel estimation. The proposed scheme computes the time-domain channel response from the received pilots. The noise power level is obtained based on the time-domain channel response by averaging the received power over a predetermined window. A novel way is further proposed to use frequency-domain mask of pilots to reduce the estimation error of the noise level. By analysis and simulation, the performance of the proposed scheme is evaluated.