Compressive Sensing Based Time Domain Synchronous OFDM with High Spectral Efficiency and Reliable Performance

From 01-06-2012 to 31-05-2015

Description

On December 6, 2011, digital terrestrial multimedia/television broadcasting (DTMB) proposed by Tsinghua university, China, is officially approved by ITU as the international digital television terrestrial broadcasting (DTTB) standard. The key technology of DTMB is time domain synchronous OFDM (TDS-OFDM), whereby the time-domain known pseudorandom-noise (PN) sequence instead of cyclic prefix (CP) is used as the guard interval of the followed OFDM data block. The known PN sequence can be also used for channel estimation and synchronization, so TDS-OFDM could achieve higher spectral efficiency and faster synchronization than standard CP-OFDM.

However, for TDS-OFDM, channel estimation and data detection has to be iteratively implemented to remove the inter-block-interference (IBI) between the PN sequence and the unknown OFDM data block, leading to the difficulty to support high-order modulation such as 256QAM where high signal-to-noise ratio (SNR) is required, as well as the obvious performance loss over doubly selective fading channels.

In this project, we will exploit the new ground-breaking compressive sensing (CS) theory fundamentally different from the classical Shannon-Nyquist sampling theorem to solve those two open problems of TDS-OFDM. Unlike the conventional TDS-OFDM scheme trying to remove IBI completely, this proposal will investigate the CS-based TDS-OFDM scheme without any interference cancellation for channel estimation. This is achieved by using a small portion of the IBI-free received PN sequence to recover the high-dimension multipath channel via the modelled channel estimation with the CS signal recovery algorithm, whereby partial priori information of the channel available in TDS-OFDM is exploited to reduce the complexity of classical signal recovery algorithm in the CS literature. In this way, channel estimation and data detection are decoupled in the CS-based TDS-OFDM scheme, and reliable yet low-complexity channel estimation can be achieved. Consequently, without changing the current infrastructure or reducing the spectral efficiency, the CS-based TDS-OFDM could support 256QAM and achieve improved performance over doubly selective fading channels. 

Team

• Marc Moonen, Promoter
• Paschalis Tsiaflakis, Team member

Financing

Funding: OTHER - Other Funding Agencies

Program/Grant Type: BIL - Bilateral Scientific Cooperation