sensors

Article
SPICE-ML Algorithm for Direction-of-Arrival
Estimation
Yu Zheng 1,2 , Lutao Liu 1, * and Xudong Yang 1
1 College of Information and Communication Engineering, Harbin Engineering University,
Harbin 150001, China; [email protected] (Y.Z.); [email protected] (X.Y.)
2 The 14th Research Institute Electronics Technology Group Corporation, Nanjing 210000, China
* Correspondence: [email protected]

Received: 29 November 2019 ; Accepted: 20 December 2019 ; Published: 24 December 2019

Abstract: Sparse iterative covariance-based estimation, an iterative direction-of-arrival approach

based on covariance fitting criterion, can simultaneously estimate the angle and power of incident
signal. However, the signal power estimated by sparse iterative covariance-based estimation approach
is inaccurate, and the estimation performance is limited to direction grid. To solve the problem above,
an algorithm combing the sparse iterative covariance-based estimation approach and maximum
likelihood estimation is proposed. The signal power estimated by sparse iterative covariance-based
estimation approach is corrected by a new iterative process based on the asymptotically minimum
variance criterion. In addition, a refinement procedure is derived by minimizing a maximum
likelihood function to overcome the estimation accuracy limitation imposed by direction grid.
Simulation results verify the effectiveness of the proposed algorithm. Compared with sparse iterative
covariance-based estimation approach, the proposed algorithm can achieve more accurate signal
power and improved estimation performance.

Keywords: direction-of-arrival estimation; covariance fitting criterion; sparse iterative covariance-based

estimation approach; maximum likelihood estimation

1. Introduction
Direction-of-arrival (DOA) refers to the process of retrieving the direction information of several
electromagnetic waves from the outputs of a number of receiving antennas that form a sensor array.
DOA estimation is a major topic in array signal processing and plays a very important role in many
applications, e.g., wireless communications, medical imaging, and radar systems [1–3],. Moreover,
wireless sensor networks (WSNs) are an emerging paradigm in wireless communications [4]. In WSNs
one of the most significant challenges is localization in which DOA is a major parameter to estimate [5,6].
The research on DOA estimation is mainly divided into two aspects: (1) array geometry, such as the
recently proposed nested array [7] and coprime array [8,9]. (2) DOA estimation algorithm with higher
estimation accuracy. In general, the algorithm applied to DOA estimation may be classified into three
parts: nonparametric, parametric, and semiparametric [10].
A main example of the nonparametric methods is delay-and-sum (DAS) beamformer [10] in
which the received signal from each sensor are weighted and delayed so as to focus on different points
in space. It is a data-independent estimation technique which is traditionally adopted due to its low
computational burden and high-signal-to-noise ratio (SNR) properties. However, data-independent
approaches suffer from leakage problem. The local leakage will reduce the resolution, which makes
the DAS beamformer unable to distinguish the incident signals with close frequency components.
Global leakage will lead to false alarms. Adaptive Capon beamformer can improve the DAS method,
but it is limited to independent signal [11]. Parametric methods, especially subspace methods such as
multiple signal classification (MUSIC) [12], estimation of signal parameters via rotational invariance

Sensors 2020, 20, 119; doi:10.3390/s20010119 www.mdpi.com/journal/sensors

Sensors 2020, 20, 119 2 of 12

techniques (ESPRIT) [13], where parameters are estimated by studying the subspace of the data
covariance. They can provide higher resolution than nonparametric methods, but adimit two
shortcomings: (1) they require the knowledge of the number of source, and (2) work well only when the
sources are independent. Recently, these two kinds of approaches have been combined, leading to the
so-called semiparametric approache. The recently developed iterative adaptive algorithm (IAA) [14]
largely eliminates the leakage problem of the beamformer and is robust to coherent sources. However,
the IAA algorithm still has resolution limitations, especially at low SNR situations. Stoica et al. have
recently proposed a user parameter-free sparse iterative covariance-based estimation (SPICE) approach
in [15,16], based on minimizing a covariance fitting criterion. SPICE is able to provide excellent
resolution and low sidelobe levels while maintaining robustness to coherent sources. Moreover, the
power of the signal can be estimated simultaneously. However, as with most power-based sparse
methods, the estimation accuracy of the SPICE algorithm is limited to the direction grid. In addition,
the signal power estimated by SPICE is inaccurate, especially in the case of coherent signal. In [17],
a method named likelihood-based estimation of sparse parameters (LIKES) is proposed, where the
same covariance fitting criterion is adopted as SPICE algorithm. The difference is that the weights in
LIKES are adaptive rather than constant in SPICE. Compared with the SPICE algorithm, LIKES can
obtain more accurate DOA estimation, but at the expense of computational cost. The DOA estimation
accuracy will be limited to direction grid because the spatial domain is also discretized in LIKES
algorithm. In [18], an iterative algorithm based on the combination of Capon and SPICE (C-SPICE) is
proposed, and a mobile average initialization technology is introduced to realize DOA estimation by
using the spatial spectrum information estimated in the previous snapshot in the next snapshot.
DAS has low computational burden and accurate signal power estimation, but low resolution.
The resolution of IAA is improved, but still not high enough to distinguish two signals which are close
to each other. SPICE has high resolution, but the signal power estimated by SPICE is not accurate, and
the DOA estimation accuracy is limited to the direction grid. In order to obtain an algorithm with
high resolution, accurate signal power estimation and high estimation accuracy, we propose a SPICE
and maximum likelihood (ML) estimation approach (SPICE-ML) in this paper. An iterative correction
process is derived through the asymptotically minimum variance (AMV) criterion [19,20] so that the
power of the signal estimated are more accurate. Moreover, a maximum likelihood cost function is
used to refine the DOA estimation, thereby combating the limit of estimation accuracy caused by
the direction grid. Numerical simulations are designed to show the effectiveness and superiority in
estimation accuracy of the proposed algorithm.
This paper is organized as follows. The signal model for DOA estimation is introduced in
Section 2 for far-field, narrowband sources. The concept of sparse representation is also discussed in
Section 2. The SPICE algorithm is introduced in Section 3, including the covariance fitting criterion and
updating formulas. Section 4 is the main highlight of this paper in which the SPICE-ML is proposed.
Compared with SPICE algorithm, SPICE-ML can receive more accurate signal power estimation and
DOA estimation accuracy. Numerical simulations are presented in Section 5, and the conclusion and
some future research directions are drawn in Section 6.
Notations: Vectors, matrices are denoted by boldface lowercase letters and boldface uppercase
letters, respectively. (·)T , (·)∗ , (·)−1 and (·) H represent the transpose, the conjugate, the inverse and the
complex conjugate transpose of vectors or matrices. ⊗ is the Kronecker product and tr (·) represents
the trace of a matrix. E [·], det (·) and vec (·) denote the operator of expectation, determinant and
vectorization. k·k denotes the Euclidean norm for vectors and the Frobenius norm for matrices.

2. Signal Model
Assume N far-field, uncorrelated narrowband signals from directions Ω= [θ1 , . . . θ N ] impinging
on an uniform line array (ULA), the number of ULA is M and the inter-element spacing equals to
a half of the signal wavelength. Let {θk }kK=1 denote a direction grid that covers Ω, where K  N,
Sensors 2020, 20, 119 3 of 12

θk ∈ [−90◦ 90◦ ). Assume the direction gird is fine enough so that the true location parameters of the
sources lie on (or, practically, close to) the grid. The signal model is shown in Figure 1.

Figure 1. Signal model.

Then the output of the array can be modeled as:

K
x (t) = ∑ a ( θ k )sk ( t ) + n ( t ) = A ( θ ) S ( t ) + n ( t ) , t = 1, . . . , T (1)
k =1

where A (θ ) = [a (θ1 ) , a (θ2 ) , . . . , a (θK )] denotes the array manifold matrix and a (θk ) is the steering
vector corresponding to θk . T denotes the number of snapshots. S (t) =[s1 (t) , s2 (t) , . . . , sK (t)]T
contains the K unknown complex-valued signals, and n (t) is the additional noise term. We assume
that E n (t) n H (t) = σI M , where σ is the power of noise, and I M is a M × M identical matrix.
 

Let us further assume that E S (t) S H (t) = Pk I M , PK = Diag ( p1 , p2 , . . . pK ), where pk represents the

unknown signal power at θk . Moreover the signal S (t) and noise n (t) are assumed to be statistically
independent. Therefore, we have the covariance matrix of x (t)

R = APK A H + σI M . (2)

In practice, this covariance matrix is usually estimated by the sample covariance matrix

T
1
R̂ =
T ∑ x ( t )x H ( t ) , (3)
t =1

Note that only few sources exist in practice, therefore in the signal matrix

··· ···
 
s1 (1) s1 (2) s1 ( T )
 .. .. .. .. ..
S= . , (4)

. . . .
s K (1) s K (2) ··· · · · sK ( T )

only a small number of rows are different from zeros. Thus, the DOA estimation problem is then
changed to decide from x (t) which rows of the signal matrix S are non-zero.

3. The SPICE Algorithm

The SPICE algorithm is derived from a robust covariance fitting criterion. It has a sound statistical
foundation, it does not require any hyper-parameters, and yet it has global convergence properties.
Moreover, the signal and noise power are estimated in a natural manner. Although the signal are
assumed to be independent, the SPICE algorithm is still robust to coherent signal [15]. In this paper,
Sensors 2020, 20, 119 4 of 12

we only consider the case in which T > M and the variances of noise are identical (see [15] for the case
of T < M and noise with different variances).
Equation (2) can be rewrite as
H
R = Aaug PAaug , (5)

where
Aaug = [a (θ1 ) a (θ2 ) . . . a (θK ) I M ] = [a1 a2 . . . aK aK +1 . . . aK + M ] , (6)

   
p1 0 ··· ··· ··· ··· 0 p1 0 ··· ··· ··· ··· 0
.. .. . . .. .. .. .. .. . . .. .. ..
0 . . ..
   

 . . . . . . .  
  . . . . 

 0 · · · · · · pK · · · · · · 0   0 · · · · · · pK ··· ··· 0 
P= = . (7)
   
 0 · · · · · · · · · σ1 · · · 0   0 ··· ··· ··· p K +1 ··· 0 
.. .. .. .. .. . . .. .. .. .. .. .. .. ..
   
.. .
   
 . . . . .   . . . . . . 
0 · · · · · · · · · · · · · · · σM 0 ··· ··· ··· · · · · · · pK + M
The SPICE algorithm considers the following weighted covariance fitting criterion:
2
f = R−1/2 R̂ − R R̂−1/2


, (8)

a simple calculation shows that Equation (8) can be expressed as:

f = tr R−1 R̂ − R R̂−1 R̂ − R




= tr R R̂ − I I − R−1 R̂
 −1


, (9)
− 1 − 1



= tr R R̂ + tr R̂ R − 2M
where

  K+ M
tr R̂−1 R = ∑ pk a∗k R̂−1 ak . (10)
k =1

The minimization of f obtained from Equations (9) and (10) is equivalent to the minimization of
the function

  K+ M  
g = tr R̂1/2 R−1 R̂1/2 + ∑ a∗k R̂−1 ak pk , (11)
k =1

The minimization problem with respect to pk in Equation (11) is a semidefinite program (SDP)
and therefore is a convex problem [21]. However, the calculation of Equation (11) as a SDP is very
computationally intensive. It can be seen from Equation (10) that a consistent estimation tr R̂−1 R = M

can be obtained when T tends to be infinite. Hence, the probem of minimizing g can be reformulated
as the following constrained minimization problem:

  K+ M
min tr R̂1/2 R−1 R̂1/2 s.t. ∑ ωk pk = 1, (12)
{ p k ≥0} k =1

where

ωk = a∗k R̂−1 ak /M. (13)

The minimization of the objective in Equation (12) can be solved by means of a cyclic
algorithm [15], leading to the updated formulas of the SPICE algorithm

a∗k R−1 (i ) R̂1/2

pik+1 = pik k = 1, . . . , K, (14)
ωk1/2 ρ (i )
Sensors 2020, 20, 119 5 of 12

K
ρ (i ) = ∑ ωk1/2 pik a∗k R−1 (i ) R̂1/2 + γ1/2 σi R−1 (i ) R̂1/2 , (15)
k =1

R−1 (i ) R̂1/2 K+ M
σ i +1 = σ i
γ1/2 ρ (i )
,γ = ∑ ωk . (16)
k = K +1

where index i denotes the number of iteration. The algorithm can be initialized by means of the DAS
method. The cyclic operation of the cyclic algorithm makes the objective function in Formula (12)
monotonically decrease, and the minimization process is a convex problem. Therefore, the result of
SPICE algorithm has global convergence. In [22], it is proved that the limit point of the iterative process
of SPICE is the global solution of the minimization problem (12) under the weak condition where
p0k > 0 and R(i ) ≥ 0 in each iteration.

4. The SPICE-ML Algorithm

SPICE algorithm has a bias in estimating the signal power, especially in the case of coherent
sources (as shown in the simulation part). To solve this problem, an iteration calibration process of
signal power is obtained by using AMV criterion. After applying the vectorization operator to the
matrix R, we can receive the vector r (p)

r (p) = vec (R) = Ãp, (17)

where à = Ã1 , āK +1 , Ã1 = [ā1 , . . . , āK ] , āk = a∗k ⊗ ak , k = 1, . . . K. āK +1 = vec (I M ) , p = [ p1 , . . . , pK ].
 
def

Note that the Gaussian circular asymptotic covariance matrix r̂ = vec R̂ is give by [20]

Cr = R∗ ⊗ R. (18)

Suppose that p can be identifiable from r (p). According to [19,20], it can be proved that the
Alg
covariance matrix Covp is bounded below by the following real symmetric positive definite matrix
h i −1
Alg
Covp ≥ ÃdH Cr−1 Ãd , (19)

where Ãd = dr (p) /dp, d (·) /d (·) is the operation of differential. In addition, we can obtain this
lower bound by minimizing the following AMV criterion

p̂ = arg min f (p)

p
def
(20)
f (p) = [r̂−r (p)] H Cr−1 [r̂−r (p)] .
From Equations (18) and (20), the estimation of p and σ are given by the following iterative s

( i +1) akH R−1(i) R̂R−1(i) ak (i ) 1

p̂k = 2
+ pk − H −1(i) , k = 1, . . . K, (21)
akH R−1(i) ak ak R ak

        
σ̂(i+1) = tr R−2(i) R̂ + σ̂(i) tr R−2(i) − tr R−1(i) /tr R−2(i) , (22)

(0) K
  n o
(i ) (i )
where R(i) = AP(i) A H + σ̂(i) I M , P(i) = Diag p̂1 , . . . , p̂K . The initialization of p̂k , σ̂(0) is
k =1
provided by the SPICE algorithm. In practice the noise power σk , k = 1, . . . M estimated by the SPICE
algorithm may be different. [23] shows that the degradation of accuracy comapared with that achieved
by imposing σk2 6= σ2 , ∀k, is not significant. Through the iterative Formula (21), the signal power are
continuously corrected.
Sensors 2020, 20, 119 6 of 12

The estimation accuracy of the SPICE algorithm is limited to the fineness of the direction grid.
A coarse grid would lead to the degradation of estimation accuracy and a problem of computational
complexity arises if a high-density grid is employed. Therefore, how to choose the direction grid
becomes a difficult problem. In order to overcome the limit of estimation accuracy caused by the
direction grid, the DOA estimation Ω̂ = θ̂1 . . . θ̂ N is refined by iteratively minimizing a stochastic

ML cost function.
The stochastic negative log-likelihood function of x (t) can be expressed as [24]
 
L (θ ) = ln (det (R)) + tr R−1 R̂ . (23)

The covariance matrix of noise is defined as

Qk = R − pk ak akH , k = 1, . . . , K. (24)

Applying matrix inverse lemma to Formula (24)

R−1 = Q− 1 H
k − pk β k bk bk , k = 1, . . . , K. (25)
  −1
where bk = Q− 1 H −1
k ak , β k = 1 + p k ak Qk ak .
Besides,    
tr R−1 R̂ = tr Q−1 R̂ − pk β k bkH R̂bk . (26)

Then use the algebraic identity det (I + AB) = det (I + BA) and obtain
h  i
ln (det (R)) = ln det Qk + pk ak akH 1 + pk akH Q− 1



= ln k ak det (Qk ) . (27)
= ln (det (Qk )) − ln ( β k )

From Equations (23), (26) and (27), we can derive

 
L (θ ) = ln (det (Qk )) + tr Q− 1
− ln ( β k ) + pk β k bkH R̂bk



k R̂
, (28)
= L (θ−k ) − l (θk )
  N
1 α2,k
l (θk ) = ln + pk , (29)
1 + pk α1,k 1 + pk + α1,k
 −1
−1

where α1,k =akH Q− 1 N H −1
k ak , α2,k = ak Qk R̂Qk ak , Qk =R − pk ak akH , k = 1, . . . , K.
The ML cost function can be decomposed into two parts: L (θ−k ), the marginal likelihood function
with parameter θk excluded, and l (θk ) concerning θk . Therefore, the ML cost function (23) with respect
to θk is equivalent to the ML function (29). Although the calculation of multi-dimensional optimization
problem is very complex, the minimization problem in Equation (29) can be effectively solved by using
the Nelder–Mead algorithm [25]. Moreover, the Nelder–Mead algorithm has already been built into
the “fminsearch” function in MATLAB. Suppose that we have obtained the corrective parameters
(i ) K (0) N
n o n o
p̂k , σ̂(i) and θ̂k , then the refinement result can be achieved by minimizing Formula (29).
k =1 k =1

The SPICE-ML algorithm is summarized in Table 1.

Sensors 2020, 20, 119 7 of 12

Table 1. The SPICE-ML algorithm.

(0) K (0) N
n o n o
step1. Initialization: p̂k , σ̂(0) and θ̂k are obtained from SPICE algorithm.
k =1 k =1
repeat
(i )
step2. Compute R(i) , Qk .
step3. Update pk and σ according to Formula (21) and (22).
 N
step4. Minimizing the ML cost function with respect to θk and obtain the refined estimated DOA θ̂k k =1 .
end for
until (convergence)

5. Simulation Results
We evaluated the performance of the proposed SPICE-ML algorithm and compared it with
DAS, IAA, and SPICE in this section. We employed a uniform linear array with 12 sensors, and the
inter-element spacing was half-wavelength. The direction grid {θk }kK=1 uniformly covered the entire
DOA space Ω = [−90◦ 90◦ ) with a step size of 0.5◦ . The SNR was defined as
p 
avg
SNR=10log10 [dB] (30)
σ
where p avg denotes the average power of the signal. The iteration termination condition is set to
p̂i+1 − p̂i 2 / p̂i 2 < 10−3 .
Three signals with power 10 dB, 8 dB, and 5 dB from θ1 = −45.3◦ , θ2 = −38.7◦ , θ3 = 30.8◦
impinged on the array. The additional noise was Gaussian white noise with equal power, and the SNR
is set to 20 dB. The number of snapshots was given by T = 50. Both independent and coherent sources
are shown in Figures 2 and 3. For the case of coherent sources, the signals at θ1 and θ3 shared the same
phases but were independent to the signal at θ2 .
It can be seen from the simulation results of Figures 2a and 3a, the DAS method failed to seperate
the two close sources at θ1 and θ2 owing to the smearing effects and limited resolution. The IAA
algorithm could significantly reduce the smearing effect so that the sidelobes in Figures 2b and 3b were
lower, but the resolution was still not high. The SPICE algorithm was capable of resolving the three
sources, but the signal power estimated was not accurate, especially in the case of coherent sources, as
shown in Figures 2c and 3c. The SPICE-ML algorithm proposed in this paper further corrected the
signal power estimated by the SPICE algorithm, leading to a more accurate signal power estimation.
Figures 2d and 3d illustrate the effectiveness and high resolution of the proposed SPICE-ML algorithm.
Next, we evaluate the root mean square error (RMSE) of DOA estimation through Monte-Carlo
simulations. The defination of RMSE is:
v
L K
u
u 1
RMSE = t
LK ∑ ∑ (θ̂k ,l −θk )2 , (31)
l =1 k =1

where θ̂k ,l is the estimated DOA of the k signal in the l-th Monte-Carlo trial, and L is the total number of
Monte-Carlo trials. Assume that three signals are randomly located at [−60◦ 60◦ ). Considering both
independent and coherent sources, respectively. The number of Monte-Carlo trials is 500. The RMSE
curves are shown in Figures 4 and 5. Moreover, we display the values of RMSE at each SNR and
snapshot in Tables 2–5.
From Figure 4a,b, it can be seen that in the case of independent signal the RMSE of all algorithms
decreased with the increase of SNR and snapshots, and the RMSE of SPICE-ML proposed in this
paper was the lowest, which illustrates that the proposed SPICE-ML approach outperformed the
other methods. From the Tables 2 and 3, specifically, compared with SPICE, SPICE-ML could
efficiently reduce the RMSE vs. SNR from 0.147◦ –0.330◦ to 0.022◦ –0.238◦ , the RMSE vs. snapshots
from 0.065◦ –0.150◦ to 0.051◦ –0.113◦ in the case of independent signal. From Figure 5a,b, we can see
the estimation accuracy of the SPICE algorithm declined dramatically in the case of coherent signal
Sensors 2020, 20, 119 8 of 12

compared with that in the case of independent signal. However, source coherence did not degrade
the DAS and IAA algorithm. Notably, SPICE-ML still offered the best estimation performance in the
case of coherent signal, because the angle estimated by SPICE was further refined by minimizing an
ML function. From the Tables 4 and 5, we can see that compared with SPICE, the RMSE vs. SNR
and snapshots were reduced from 0.195◦ –2.777◦ to 0.023◦ –0.244◦ , from 0.276◦ –1.950◦ to 0.053◦ –0.114◦ ,
respectively. We can also observe that there existed plateau effects for DAS, IAA, and SPICE. This is
because all these methods estimate DOA by means of selecting one element from a fixed set of direction
grid, but there always exists an estimation bias no matter how fine the direction grid is. Theoretically,
bias can be reduced by selecting a dense gird. However, a dense grid will lead to large computation
cost and is not applicable in practice. On the contrary, the SPICE-ML algorithm would not suffer
from the plateau effect because of a refinement DOA estimation procedure based on minimizing a ML
cost function.

25 25
DAS IAA
20 20
True DOA True DOA

15 15

10 10

5 5
Power (dB)

Power (dB)
0 0

-5 -5

-10 -10

-15 -15

-20 -20

-25 -25

-30 -30
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
°) °)

(a) DAS algorithm (b) IAA algorithm

25 25
SPICE-ML
SPICE
20 20 True DOA
True DOA

15 15

10 10

5 5
Power(dB)

0 0

-5 -5

-10 -10

-15 -15

-20 -20

-25 -25

-30 -30
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
°)

(c) SPICE algorithm (d) SPICE-ML algorithm

Figure 2. Root mean square error (RMSE) with independent signal.

Sensors 2020, 20, 119 9 of 12

25 25
DAS IAA
20 20
True DOA True DOA

15 15

10 10

5 5
Power (dB)

Power (dB)
0 0

-5 -5

-10 -10

-15 -15

-20 -20

-25 -25

-30 -30
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
°) °)

(a) DAS algorithm (b) IAA algorithm

25 25
SPICE SPICE-ML
20 20
True DOA True DOA

15 15

10 10

5 5
Power (dB)

0 0

-5 -5

-10 -10

-15 -15

-20 -20

-25 -25

-30 -30
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
°)

(c) SPICE algorithm (d) SPICE-ML algorithm

Figure 3. RMSE with coherent signal.

10 0 10 0
DAS DAS
IAA IAA
SPICE SPICE
SPICE-ML SPICE-ML
RMSE/°
RMSE/°

10 -1 10 -1

10 -2 10 -2
0 2 4 6 8 10 12 14 16 18 20 20 30 40 50 60 70 80 90 100
SNR/dB snapshots

(a) RMSE vs. SNR with 50 snapshots (b) RMSE vs. snapshots with 10 dB SNR

Figure 4. RMSE with independent signal.

Sensors 2020, 20, 119 10 of 12

10 1 10 1
DAS DAS
IAA IAA
SPICE SPICE
SPICE-ML SPICE-ML

10 0 10 0
RMSE/°

RMSE/°
10 -1 10 -1

10 -2 10 -2
0 2 4 6 8 10 12 14 16 18 20 20 30 40 50 60 70 80 90 100
SNR/dB snapshots

(a) RMSE vs. SNR with 50 snapshots (b) RMSE vs. snapshots with 10 dB SNR

Figure 5. RMSE with coherent signal.

Table 2. RMSE vs. signal to noise ratio (SNR) with independent signal.

SNR 0 2 4 6 8 10 12 14 16 18 20
DAS 0.339 0.315 0.285 0.274 0.271 0.262 0.258 0.260 0.253 0.258 0.249
IAA 0.365 0.297 0.250 0.226 0.198 0.180 0.170 0.161 0.156 0.153 0.149
SPICE 0.330 0.263 0.223 0.192 0.177 0.166 0.158 0.152 0.149 0.148 0.147
SPICE-ML 0.238 0.187 0.144 0.112 0.088 0.070 0.057 0.044 0.035 0.028 0.022

Table 3. RMSE vs. snapshots with independent signal.

snapshots 20 30 40 50 60 70 80 90 100
DAS 0.342 0.257 0.241 0.241 0.209 0.196 0.184 0.180 0.178
IAA 0.148 0.128 0.125 0.112 0.110 0.105 0.100 0.102 0.102
SPICE 0.150 0.105 0.095 0.081 0.078 0.073 0.068 0.063 0.065
SPICE-ML 0.113 0.089 0.082 0.069 0.064 0.061 0.055 0.051 0.051

Table 4. RMSE vs. SNR with coherent signal.

SNR 0 2 4 6 8 10 12 14 16 18 20
DAS 0.625 0.617 0.593 0.589 0.589 0.590 0.588 0.582 0.579 0.581 0.579
IAA 0.376 0.322 0.273 0.233 0.210 0.190 0.179 0.166 0.159 0.155 0.150
SPICE 2.777 0.811 0.777 0.407 0.360 0.312 0.275 0.252 0.226 0.199 0.195
SPICE-ML 0.244 0.187 0.147 0.110 0.091 0.069 0.055 0.045 0.034 0.028 0.023

Table 5. RMSE vs. snapshots with coherent signal.

snapshots 20 30 40 50 60 70 80 90 100
DAS 0.637 0.626 0.617 0.613 0.617 0.606 0.603 0.609 0.609
IAA 0.163 0.143 0.135 0.130 0.124 0.124 0.117 0.112 0.117
SPICE 1.950 0.364 0.321 0.311 0.297 0.285 0.285 0.292 0.276
SPICE-ML 0.114 0.094 0.078 0.074 0.065 0.062 0.056 0.052 0.053

6. Conclusions
In this paper, we combined the SPICE DOA estimation algorithm and maximum likelihood
estimation and proposed a SPICE-ML algorithm. Compared with the SPICE algorithm, the SPICE-ML
method derive an iterative correction procedure for signal power estimation based on the AMV
criterion, and combat the limitation of the direction grid by minimazing a Maximum Likelihood
cost function. The simulation results verify the effectiveness of the proposed algorithm. In addition,
Sensors 2020, 20, 119 11 of 12

the superiority of the SPICE-ML algorithm in DOA estimation accuracy is illustrated compared
with the DAS, IAA, and SPICE algorithms. Specifically, comparing with SPICE, SPICE-ML can
efficiently reduce the RMSE vs. SNR from 0.147◦ –0.330◦ to 0.022◦ –0.238◦ in the case of independent
signal, from 0.195◦ –2.777◦ to 0.023◦ –0.244◦ in coherent signal situation, the RMSE vs. snapshots from
0.065◦ –0.150◦ to 0.051◦ –0.113◦ in the case of independent signal, from 0.276◦ –1.950◦ to 0.053◦ –0.114◦
in coherent signal situation. In the future, it will be of interest to study the performance of SPICE-ML
algorithm in the case of single snapshot. This paper only considers one-dimensional DOA estimation,
so another future direction of interest would be the extension of SPICE-ML to two-dimensional
DOA estimation.

Author Contributions: L.L. and Y.Z. proposed the original idea. Y.Z. and X.Y. designed the simulations and
provided all of the figures and data. Y.Z. wrote the paper under the guidance of L.L., and X.Y. checked the paper.
All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded in part by the National Natural Science Foundation of China under
Grant 61571146, and in part by the Fundamental Research Funds for the Central Universities under Grant
HEUCFP201908.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Gorcin, A.; Arslan, H. A two-antenna single RF front-end DOA estimation system for wireless
communications signals. IEEE Trans. Antennas Propag. 2014, 62, 5321–5333. [CrossRef]
2. Synnevag, J.F.; Austeng, A.; Holm, S. Adaptive beamforming applied to medical ultrasound imaging.
IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2007, 54, 1606–1613. [CrossRef] [PubMed]
3. Yao, B.; Wang, W.; Yin, Q. DOD and DOA estimation in bistatic non-uniform multiple-input multiple-output
radar systems. IEEE Commun. Lett. 2012, 16, 1796–1799. [CrossRef]
4. Strumberger, I.; Minovic, M.; Tuba, M.; Bacanin, N. Performance of elephant herding optimization and
tree growth algorithm adapted for node localization in wireless sensor networks. Sensors 2019, 19, 2515.
[CrossRef]
5. Paul, A.K.; Sato, T. Localization in wireless sensor networks: A survey on algorithms, measurement
techniques, applications and challenges. J. Sens. Actua. Netw. 2017, 6, 24. [CrossRef]
6. Huang, S.C.; Li, F.G. A novel positioning system based on coverage area pruning in wireless sensor networks.
Sensors 2018, 18, 4469. [CrossRef]
7. Pal, P.; Vaidyanathan, P.P. Nested arrays: A novel approach to array processing with enhanced degrees of
freedom. IEEE Trans. Signal Process. 2010, 58, 4167–4181. [CrossRef]
8. Zhou, C.; Gu, Y.; Fan, X.; Shi, Z.; Mao, G.; Zhang, Y.D. Direction-of-arrival estimation for coprime array via
virtual array interpolation. IEEE Trans. Signal Process. 2018, 66, 5956–5971. [CrossRef]
9. Zhou, C.; Gu, Y.; Shi, Z.; Zhang, Y.D. Off-Grid direction-of-arrival estimation using coprime array
interpolation. IEEE Signal Process Lett. 2018, 25, 1710–1714. [CrossRef]
10. Stoica, P. SPECTRAL ANALYSIS OF SIGNALS: The missing data case. Synth. Lect. Signal Process. 2016, 1,
90–108.
11. Liu, H.; Liao, G.; Zhang, J. A robust adaptive Capon beamforming. Signal Process. 2006, 86, 2820–2826.
12. Schmidt, R.; Schmidt, R.O. Multiple emitter location and signal parameters estimation. IEEE Trans.
Antennas Propag. 1986, 34, 276–280. [CrossRef]
13. Roy, R.; Kailath, T. ESPRIT-Estimation of signal parameters via rotational invariance techniques. IEEE Trans.
Acoust. Speech Signal Process. 1989, 37, 984–995. [CrossRef]
14. Yardibi, T.; Li, J.; Stoica, P. Source localization and sensing: A nonparametric iterative adaptive approach
based onweighted least squares. IEEE Trans. Aerosp. Electron. Syst. 2010, 46, 425–443. [CrossRef]
15. Stoica, P.; Babu, P.; Li, J. SPICE: A sparse covariance-based estimation method for array processing.
IEEE Trans. Signal Process. 2011, 59, 629–638. [CrossRef]
16. Stoica, P.; Babu, P.; Jian, L. New method of sparse parameter estimation in separable models and its use for
spectral analysis of irregularly sampled data. IEEE Trans. Signal Process. 2010, 59, 35–47. [CrossRef]
Sensors 2020, 20, 119 12 of 12

17. Stoica, P.; Babu, P. SPICE and LIKES: Two hyperparameter-free methods for sparse-parameter estimation.
Signal Process. 2012, 92, 1580–1590. [CrossRef]
18. Cai, S.; Shi, X.; Zhu, H. Direction-of-arrival estimation and tracking based on a aequential implementation of
C-SPICE with an Off-Grid model. Sensors 2017, 17, 2718. [CrossRef]
19. Abeida, H.; Delmas, J.P. Efficiency of subspace-based DOA estimators. Signal Process. 2007, 87, 2075–2084.
[CrossRef]
20. Delmas, J.P. Asymptotically minimum variance second-order estimation for noncircular signals with
application to DOA estimation. IEEE Trans. Signal Process. 2004, 52, 1235–1241. [CrossRef]
21. Yurii, N. Lectures on convex optimization. IEEE Trans. Autom. Control 2006, 51, 1859–1859.
22. Yu, Y. Monotonic convergence of a general algorithm for computing optimal designs. Ann. Stat. 2010, 38,
1593–1606. [CrossRef]
23. Stoica, P.; Zachariah, D.; Li, J. Weighted SPICE: A unifying approach for hyperparameter-free sparse
estimation. Digital Signal Process. 2014, 33, 1–12. [CrossRef]
24. Stoica, P.; Nehorai, A. Performance study of conditional and unconditional direction-of-arrival estimation.
IEEE Trans. Acoust. Speech Signal Process. 1990, 38, 1783–1795. [CrossRef]
25. Nelder, J.A.; Mead, R. A simplex method for function minimization. Commput J. 1965, 7, 308–313. [CrossRef]

c 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).