Sparsity-aware direct decision-feedback equalization of ionospheric HF channels

We compare a compressive sensing-based sparse direct decision-feedback equalization with the conventional structure in a MIL-STD-188-110C military HF receiver. The Stochastic Gradient Pursuit (SGP) adaptive algorithm is considered for the sparse DFE approach, while the conventional DFE runs a variation of the well-known Recursive Least Squares (RLS) algorithm. In addition, a variant of the SGP algorithm (κ-SGP) with reduced support set computations is proposed and analyzed. Results show that the sparsity of the DFE filters can also be exploited in the context of HF communications to derive efficient and better performing receiver implementations.