Development of an Efficient Low-complexity Channel Estimator for Digital Television Terrestrial Broadcasting Systems

Orthogonal frequency division multiplexing (OFDM) is widely used to transmit data in many wireless communication applications including digital television terrestrial broadcasting (DTTB). Although the existing dual pseudo noise padding (DPNP) based time-domain synchronous OFDM (TDS-OFDM) system has low complexity, the spectral efficiency is low. The time-frequency-domain (TFD) based frame structure enhances the system performance of TDS-OFDM over fast time-varying channels by compromising with computational complexity. This paper proposes a novel frame structure for OFDM-based DTTB system which incorporates pilots in the time domain, and retains the cyclic prefix and modulable orthogonal sequence (MOS) from the TFD-based frame structure. Since the proposed frame structure is completely defined in time domain, channel estimation and equalization become easier. Using the new frame structure, a novel channel estimation technique is proposed that works in two stages. In the first stage, the MOS in guard interval is used to estimate the channel delay and gain. In the second stage, the channel gains estimated in first stage are fine-tuned using adaptive algorithms such as least mean square (LMS) or recursive least squares algorithm. The bit-error-rate (BER) performance of the proposed two-stage channel estimation technique is better compared to that of DPNP-based TDS-OFDM. In addition, computational complexity of the proposed LMS-based two-stage channel estimation approach is low compared to TFD-based TDS-OFDM system. Less than 1.5% of the total sub-carriers are used as redundant pilots, and therefore the loss in spectral efficiency is negligible in the proposed approach.