Compressive Sensing Based Channel Estimation for Layered Modulation OFDM Scheme

Time-domain synchronous orthogonal frequency division multiplexing (TDS-OFDM) is widely recognized as an orthogonal frequency division multiplexing (OFDM) transmission scheme with high spectral effciency. However, it suffers from severe performance loss over strong frequency selective channels, thus has difficulty supporting high-order modulations such as 256 QAM. To improve the channel estimation (CE) performance and support high-order modulation in TDS-OFDM system, we divide the OFDM block into several independent layers which are encoded and modulated separately. In the receiver, the pseudo-random noise (PN) sequence can give a coarse CE and then the low-order modulated symbols can be accurately decoded. By using the auxiliary information based subspace pursuit (A-SP) algorithm on the demodulated data, a precise CE can be given to support high-order modulated data's decoding. Simulations show that the proposed compressive sensing based channel estimation scheme performs well under different channels.