Compressive sensing based direction-of-arrival estimation using reweighted greedy block coordinate descent algorithm for ESPAR antennas

Abstract

In this paper, we consider the problem of direction-of-arrival (DoA) estimation using electronically steerable parasitic array radiator (ESPAR) antenna based on compressive sensing. For an ESPAR antenna, the beampatterns and sparse model of DoA estimation problem in terms of overcomplete dictionary and sampling grid is presented. The DoA estimation problem is formulated as a mixed-norm ℓ2,1 minimization problem and the reactance domain multiple signal classification (RD-MUSIC) spatial spectrum for ESPAR antenna is introduced. Then, we propose reweighted greedy block coordinate descent (RW-GBCD) and reweighted ℓ2,1-SVD (RW-ℓ2,1-SVD) algorithms for DOA estimation using ESPAR. The performance of RW-GBCD for DoA estimation is compared to that of GBCD, ℓ2,1-SVD and RD-MUSIC algorithms. RW-GBCD benefits from less computational complexity compared to RW-ℓ2,1-SVD. Simulation results demonstrate that the performance of RW-GBCD is better than that of GBCD and ℓ2,1-SVD. When angle separation is less than 10°, RW-ℓ2,1-SVD outperforms RW-GBCD. However, when angle separation is more than 10°, the performance of RW-GBCD in terms of root mean square error (RMSE) is approximately the same as that of RW-ℓ2,1-SVD.