Existing direction-of-arrival and polarization estimation methods mostly rely on subspace technique.This paper proposes a novel DOA and polarization angle estimation algorithm from sparse signal reconstruction perspective.The algorithm first constructs a cumulant matrix model which is only related to DOA parameter

and then obtains DOA estimation using the weighted l

?

1

-norm minimization.Further

this paper constructs another three cumulant vector models by sum-average arithmetic

and enforces sparsity by Zhang penalty

which leads to almost unbiased polarization angle estimation.Meanwhile

this paper also demonstrates how to identify two sources with same DOA using their polarization characteristics.Compu

ter simulation results validate the effectiveness of the proposed algorithm.