Narrowband Angle of Arrival Estimation Exploiting Graph Topology and Graph Signals

Abstract

Based on recent results of applying graph signal processing (GSP) to narrowband angle of arrival estimation for uniform linear arrays, we generalise the analysis to the case of arrays with elements placed arbitrarily in three dimensional space. We comment on the selection of the adjacency matrix, analyse how this new approach compares to the multiple signal classification (MUSIC) algorithm, and provide an efficient implementation. We demonstrate that the GSP approach can perform as well as the MUSIC algorithm in terms of accuracy and computational cost. Simulations indicate that the proposed GSP approach avoids the severe performance degradation with which MUSIC is associated at low signal to noise ratios.

Graph-based Matched Field Localization for an Underwater Source

“…In this way, p(ω) becomes a special eigenvector t u of the adjacency matrix A(γ 0 , ω), which corresponds to the unit eigenvalue. It is worth noting that, compared with [10,11], only a spatial adjacency matrix is used. This is due to the fact that the temporal adjacency matrix requires a precise time shift, which is difficult to satisfy in the real underwater scenarios.…”

“…In [10], the authors build a cyclic graph structure in the spatial and temporal domains, such that the array signal becomes an eigenvector of the adjacency matrix corresponding to the unit eigenvalue. [11] further develops the approach by choosing a fully connected graph and highlighting the performance difference between the GSP and MUSIC methods.…”

“…Inspired by the previous studies, in this paper we implement the GSP method into the MFP location model. Like [10] and [11], we also focus on the location problem of a single source. Due to the difficult deployment of the underwater array and the complex interactions among the sensors, the adjacency matrix is designed for the arbitrarily-distributed array based on the Green's function.…”

Joint Estimation of Azimuth and Distance for Far-Field Multi Targets Based on Graph Signal Processing

“…Graph signal processing uses a new data structure that studies the connection of things, which has shown excellent performance in many fields such as graph neural network and graph cuts [3]. Some related works can be found focusing on the use of graph signals to deal with DoA estimation problems in the radar array system [4][5][6], microphone and speakers [7,8], and the sonar array system [9]. Experiments show that the graph signal processing-based DoA methods have better performance than traditional algorithms such as Multiple Signal Classification (MUSIC) in a low signal-to-noise ratio environment [10,11].…”

“…where X ∈ C M×Q represent x(k) in (5), which is the snapshot signal samples collected by M receiving antennas; Σ = diag {σ i } I i=1 is the diagonal matrix of the RCS scattering coefficient of the far-field targets; W ∈ C M×Q is the sampling noise term of n(k) in Equation ( 5), usually additive white gaussian noise (AWGN). The receiving steering vector B includes the two-way phase delay, so the expression is the same as the steering vector st (θ i ,R i ,M) in Equation ( 4):…”