Sparse doubly-selective channels: Estimating path parameters unambiguously

Doubly selective channels or time-varying multipath channels occur when communication systems are expected to work in a highly mobile environment. The estimation and the subsequent equalization of such channels is a non-trivial task. Thus, a channel estimation scheme that is robust, precise and works with a complexity that is applicable for consumer applications is vital to overcome the effects of such a channel at the receiver. The Matching Pursuit (MP) algorithm is a Compressed Sensing (CS) scheme that is shown to work well in scenarios of high mobility. However, good performance is achieved only when the right search metric is used. In this paper, an appropriate metric for the estimation of the multipath delays is proposed. It is shown that this metric produces an accurate estimate of the multipath delays under different channel conditions. In addition to this, a novel method that implicitly estimates the Doppler shift is proposed. The results show that the proposed schemes accurately estimate a doubly selective channel as compared to the classical MP and the Least Squares (LS) channel estimate. Moreover, an implicit estimation of the Doppler shift reduces the computational burden at the receiver resulting in a lower complexity when compared to the classical MP algorithm. The proposed scheme is implemented for the IEEE 802.11p standard and is applicable in any Orthogonal Frequency-Division Multiplexing (OFDM) based wireless system that is expected to work in highly mobile environments, specifically for 5G.