Estimating Doubly-Selective Channels in DVB-T2

The ability to receive broadcast content in highly mobile scenarios plays an important role in determining the success of broadcast systems in general. The second generation of Terrestrial Digital Video Broadcasting (DVB-T2) has made several technical enhancements to increase the data carrying capacity that enabled the reception of High Definition (HD) content. Due to its properties, Orthogonal Frequency-Division Multiplexing (OFDM) is chosen as the transmission scheme in the physical layer. However, in highly mobile environments, a doubly-selective or a time-varying multipath channel that exhibits selectivity in both the time as well as the frequency domain, deteriorates the performance of the receiver. Furthermore, DVB-T2 allows up to 32k carriers on the same channel raster resulting in a smaller carrier spacing that further amplifies the effects of mobility at the receiver. Thus, channel estimation and equalization techniques play a vital role in overcoming the effects of a doubly-selective channel and enabling high mobility in a DVB-T2 receiver. Channel estimation schemes must not only be able to model the wireless channel accurately but also be computationally efficient in order to enable implementation on resource constrained consumer hardware. Compressed Sensing (CS) schemes like the Rake-Matching Pursuit (RMP) algorithm are proven to perform well in vehicular communication systems and the associated computational complexity is significantly smaller than competing state-of-art estimation schemes. In this paper, we show that the RMP algorithm for channel estimation is suitable to be used with DVB-T2 under varying channel conditions. The results show that the proposed schemes produce significantly better results when compared to the conventional channel estimation and equalization schemes. Furthermore, a detailed analysis on the computational complexity will further strengthen the need for the proposed scheme to estimate doubly-selective channels.