† Corresponding author. E-mail:
Project supported by the National Natural Science Foundation of China (Grant No. 11534009).
The inhomogenous ocean waveguide, which leads the amplitude and phase of the signal arriving at a hydrophone array to fluctuate, is one of the causes that make the array gain deviate from its ideal value. The relationship between the array gain and the fluctuant acoustic channel is studied theoretically. The analytical expression of the array gain is derived via an acoustic channel transfer function on the assumption that the ambient noise field is isotropic. The expression is expanded via the Euler formula to give an insight into the effect of the fluctuant acoustic channel on the array gain. The result demonstrates that the amplitude fluctuation of the acoustic channel transfer functions has a slight effect on the array gain; however, the uniformity of the phase difference between the weighting coefficient and the channel transfer function on all the hydrophones in the array is a major factor that leads the array gain to further deviate from its ideal value. The numerical verification is conducted in the downslope waveguide, in which the gain of a horizontal uniform linear array (HLA) with a wide-aperture operating in the continental slope area is considered. Numerical result is consistent with the theoretical analysis.
Array gain is defined as the improvement in the signal-to-noise ratio (SNR) obtained at the array output over that at a single hydrophone.[1,2] It is one of the most important measures of the sonar system performance which is affected by the ocean waveguide. The array gain reaches its ideal value 10lgM (M is the number of the hydrophones in the array) when the signals arrive as plane waves and the noises are uniform and uncorrelated between the receivers. However, the sound waves propagating in the ocean are bound to be affected by acoustic channels. And subject to the topography and acoustic properties of sea-floor, as well as eddies, tides, internal wave, and surface gravity waves, the ocean waveguides are always randomly inhomogeneous, which leads the acoustic channels to suffer from spatiotemporal variation. Besides, in the continental slope area,[3] the ocean bottom is range-dependent, which will result in the more complex acoustic channels as manifested by inhomogenous amplitude and corrugated wavefront. A distorted wavefront or the multipath effect leads the amplitude and phase of the received signals (not the plane waves) to fluctuate. Consequently, the array gain will deviate from its ideal value[4] especially for a wide-aperture uniform linear array, which ultimately fails to achieve the design performance of a sonar system in the practical application.
Earlier researches on the array gain influenced by the amplitude and phase fluctuations of the signals were based on the hypothesis that the fluctuations conform to a statistical regularity. The expression of the array gain for a linear array can be obtained if the phase fluctuation is consistent with the multidimensional normal distribution.[5–7] Kleinberg[8] investigated the effect of the amplitude and phase fluctuations on the array gain when the fluctuations were normally distributed. Using signal correlations to characterize the signal amplitude and phase fluctuations, Cox[9] and Green[10] have derived the expression of the array gain in the uncorrelated ambient noise field when the signal correlations were presented as exponential attenuation and linear attenuation, respectively. However, the amplitude and phase fluctuations and the signal correlation attenuation are affected by the ocean waveguide, which cannot always be described by the regular rules.
With the development of the underwater acoustic field modeling, the acoustic channel can be modeled by different acoustic field theories such as ray-based model, normal-mode model, wavenumber integration method, and parabolic-equation method. It provides a new approach to studying the influence of the acoustic channel fluctuation on the array gain. Recently, the majority of researches of the array gain focused on the linear array in the shallow water[11–17] and deep ocean,[18] both of which are the range-independent ocean waveguide, based on the normal-mode model.[19] In the range-independent waveguide, the signal field is characterized by a set of discrete normal modes of propagation, and the array gain can be formulated by normal modes. But in the continental slope area,[3] the acoustic wave propagation will span the different seas. In this case, the ocean waveguide is not horizontally layered, but it is dependent on the range, which leads the modes to be coupled seriously. Hence, the analytical methods of the array gain in the range-independent waveguide based on the normal-mode model are not applicable in the case of the range-dependent waveguide.
In Ref. [3], we have discussed the gain of a linear array relating to the structure of the acoustic field from the acoustic field correlation and the transmission loss. In the present paper, we will provide a new theoretical method to further analyze the array gain affected by the fluctuant acoustic channel, which can be applied to the range-dependent and range-independent waveguides. Assuming that the ambient noise is isotropic, the expression of the array gain is derived as a function of the acoustic channel transfer functions, and then the expression is expanded via the Euler formula to investigate the relationship between the array gain and the acoustic channel. The rest of this paper is organized as follows. The expression of the array gain considering the acoustic channel transfer function is derived in Section
The array gain of an arbitrary array of hydrophones is defined as
The effects of the array processing on signal and on noise can be considered separately.[12] The improvement on the signal power obtained by using an array is defined as the array signal gain, and improvement on the noise power as the array noise gain, denoted as gas = Sarray/Saverage_hyp and gan = Narray/Naverage_hyp, respectively, where Sarray and Saverage_hyp are the signal power at the array output and the average signal power of all hydrophones, Narray and Naverage_hyp are the noise power at the array output and the average noise power of all hydrophones, respectively. Then the array gain in Eq. (
The propagation of sound waves in the ocean can be described by the wave equations. For a short durative pulse, the acoustic channel can be characterized by the impulse response h(t) in the time domain or the transfer function H(f) in the frequency domain. The acoustic channel transfer functions can be calculated by solving the wave equation through utilizing different acoustic field theories[19] for range-dependent and range-independent waveguides.
Assume that the source radiates a continuous-wave (CW) signal with a frequency of f0. Then the receiving signal in the frequency domain at the i-th hydrophone can be described as
According to Parseval’s theorem, the signal energy at the i-th hydrophone is
Therefore, the array signal gain can be expressed as
The array noise gain of the array with an equal inter-element spacing d in an isotropic noise environment has been derived as[3]
Substituting Eqs. (
To investigate the relationship between the array gain and the fluctuant acoustic channels, we expand Eq. (
The acoustic channel transfer function and the weighting coefficient corresponding to the i-th hydrophone can be respectively expressed by complex numbers as
Substituting Eq. (
1) When the phase differences on all the hydrophones in the array are uniform, i.e., Δi and Δk satisfy the relationship Δi − Δk = 2nπ, where n is an arbitrary integer, equation (
In this case, the array gain only relates to the amplitudes of the channel transfer functions corresponding to each hydrophone, which means that the array gain is closely related to the fluctuation in transmission loss. According to the Cauchy–Schwarz inequality
2) When the phase differences are nonuniform, in which case Δi − Δk ≠ 2nπ, one can obtain
Generally, for an array in the ocean waveguide, the amplitude fluctuation of the acoustic channel transfer function is relatively small within the scope of the whole array. Therefore, the array gain will deviate from the ideal value slightly if it is only affected by the amplitude fluctuation. However, the array gain suffers more the phase fluctuation of the channel transfer function. Especially, if the differences between two arbitrary Δs are in the second quadrant or the third quadrant, i.e., 0.5π + 2nπ ˂ Δi − Δk ˂ π + 2nπ or −π + 2nπ ˂ Δi − Δk ˂ −0.5π + 2nπ, in the expansion of the summation formula in the numerator of Eq. (
Assuming that the ambient noise field is isotropic, we focus on investigating the relationship between the array gain and the acoustic channel transfer function in a continental downslope waveguide and also on verifying the theoretical analysis in Section
Numerical simulations are conducted in a downslope waveguide, and the results are shown in Fig.
The sound-speed profile (SSP) of the shallow water has a negative gradient as shown by the blue solid line in Fig.
In the simulation, the depth of the array is changed in a range of 10–1300 m and the array gain is calculated every 20 m; the source–receiver range is changed from 20 km to 90 km and the array gain is calculated every 200 m.
Firstly, we consider an ideal case that the weighted phases can compensate for the phase differences for all the hydrophones, and investigate the effect of amplitudes fluctuation on the array gain. Assuming that the phases of the acoustic channel transfer functions are known, we take the conjugate phases of the channel transfer functions as the phases of the weighting coefficients, i.e., θi = ψi and wi = ejψi, in which case the array gain can be expressed as Eq. (
As shown in Fig.
Next, consider a general case. We take conventional beamformer (CBF) for example to study the influence of the fluctuant acoustic channel on the array gain. The weighting vector of CBF is
The simulation parameters are the same as the above, the array gains of CBF are plotted in Fig.
Comparing Fig.
For further analyses, we choose the array gains respectively at the receiving-depth of 100 m (in the white rectangle) and 300 m (out of the white rectangle) to investigate the effect of the fluctuant acoustic channel on the array gain. We present the array gain as a function of source–receiver range changing from 25 km to 90 km. Figure
We choose the array gains when the source–receiver ranges are 36 km (in the white rectangle) and 54 km (out of the white rectangle) to analyze the effects of the fluctuant acoustic channel on the array gain. Figure
The phase differences (Δs) between the weighting coefficients of CBF and the channel transfer functions are correspondingly shown in Fig.
We also analyze the fluctuant acoustic channel and array gain when the source–receiver ranges are 36 km and 54 km (the same as the above values). The amplitudes and phases of the acoustic channel transfer functions on each hydrophone are plotted in Figs.
The corresponding array gains of CBF are 19.7 dB and 19.9 dB, which are very close to the ideal value as shown in Fig.
In addition, despite the amplitude fluctuation on the range of 36 km being larger than that on the range of 54 km, the corresponding array gain has a value almost the same as that resulting from the case of uniform Δs. It also verifies that the phase fluctuation of the acoustic channel transfer function has a main influence on the array gain.
From the analyses in this section, it concludes that the amplitude fluctuation has a slight influence on the array gain, but the uniformity of Δs is a major factor that leads the array gain to deviate from the ideal value. The uniform enhancement of Δs will increase the array gain. Furthermore, the array gain will be close to the ideal value if the difference between two arbitrary Δs approximates to 2nπ, in which case only the amplitude fluctuation reacts on the array gain. The numerical results verify the theoretical analyses in Section 3.
The array gain that deviates from the ideal value in an ocean waveguide is investigated from the fluctuant acoustic channel. The analytical expression of array gain considering the acoustic channel is derived on the assumption that the ambient noise field is isotropic. The discussion provides an insight into the physics of the problem as well as enabling quantitative estimates of the array performance in the ocean waveguide.
The fluctuant acoustic channel can be considered by the amplitude fluctuation and the phase fluctuation of the channel transfer function, then we derive the expressions for the array gain as given by Eqs. (
To verify the theoretical analysis, the gains at the depths of 100 m and 300 m in the downslope waveguide are chosen to contrastively analyze the effect of acoustic channel fluctuation on the array gain. Both the theoretical and numerical results show that the uniformity of Δ has a main influence on the array gain. Furthermore, the array gain will be close to the ideal value if the difference between two arbitrary Δs is approximately 2nπ.
[1] | |
[2] | |
[3] | |
[4] | |
[5] | |
[6] | |
[7] | |
[8] | |
[9] | |
[10] | |
[11] | |
[12] | |
[13] | |
[14] | |
[15] | |
[16] | |
[17] | |
[18] | |
[19] | |
[20] | |
[21] | |
[22] | |
[23] |