2D-DOA Estimation of Strictly Noncircular Sources Utilizing Connection-Matrix for L-Shaped Array

In wireless communication, the signals with noncircularity are widely applied. In this letter, we consider the problem of two-dimensional direction of arrival estimation with L-shaped array for strictly noncircular sources. To obtain more precise estimation with lowest possible computational complexity, we construct two connection-matrices to establish the relation between the data corresponding to the two orthogonal uniform linear arrays in L-shaped array so that we can still exploit all information of the extended received signal after dimensionality reduction. On this basis, the rank-reduction technique and one-dimensional polynomial root finding technique are able to be utilized only twice to obtain unpaired estimation results. Finally, a concise angle pairing is performed via the third connection-matrix. Compared with existing algorithms, the proposed algorithm can obtain more precise estimation and better resolution. Simulation results are provided to validate the superiority of the proposed algorithm.

I. Introduction

In wireless communication, binary phase shift keying (BPSK) and amplitude shift keying (ASK) signals are widely applied. Different from conventional signals, BPSK and ASK signals have noncircularity property, which makes them be considered as noncircular (NC) signals [1]. For direction of arrival (DOA) estimation, the degrees of freedom (DOFs) can be increased by exploiting the noncircularity of the NC signals, which improves the estimation accuracy and increases the number of resolvable sources [1]. Consequently, DOA estimation for NC sources attract lots of attention. A number of methods, such as multiple signals classification (MUSIC) [2], estimation of signal parameters via rotational invariance technique (ESPRIT) [3], and generalized ESPRIT [4], have been proposed for one-dimensional (1-D) DOA estimation for NC sources. These methods evolve from their well-known prototypes, i.e., MUSIC [5], ESPRIT [6] and generalized ESPRIT [7], proposed for circular signals.