Passive Acoustic Ranging by Multimode Waveguide Interferometry


A. Turgut, M.H. Orr, and B.H. Pasewark
 Acoustics Division

Introduction: To achieve the power projection to land mission, the Navy needs the ability to passively estimate the range of submerged targets operating in a dynamic littoral environment. The authors are investigating a passive range estimation technique that may be robust in environmentally variable littoral operating areas. The technique will require the deployment of four lightly populated vertical arrays (2 to 4 hydrophones). Acoustic signals received on the vertical arrays will be used to form "virtual horizontal arrays." Range ratios (the ratio of the distance from two vertical arrays and the target) estimated from "virtual horizontal arrays" derived from pairs of the four vertical array system will be used to localize a passive target. We have used acoustic data taken in a dynamic littoral setting to demonstrate that accurate range ratios can be estimated from two vertical line arrays (VLAs). The technique hinges on the fact that broadband acoustic signals propagate in the shallow water as modes with slightly different phase speeds and, as a result, create interference structures in both range and frequency.

Passive Acoustic Ranging in Littoral Waters: We combine two techniques to estimate the range ratio of a passive target in littoral waters. The first is to generate a "virtual receiver" by applying holographic array processing1 to broadband acoustic signals received on two VLAs. The second is the use of a waveguide invariant2 parameter to form a "virtual receiver array" from a virtual receiver. Turgut has combined the two techniques and derived an analytical formulation that describes the signal interference structure of the virtual array output.3 The formulation can be used to estimate the range ratio to a target when its signal is received simultaneously on two VLAs. The analytic formula predicts the interference structure for a given range ratio r1/r2. The range ratio is defined as the ratio of r1 and r2, i.e., the range from each vertical line array to the passive target of interest. By using the range ratios extracted from a set of four VLAs, analysis indicates that the absolute range of the target from the receivers can be determined.

Illustration: Figure 4 is a three-dimensional sketch of the geometrical configuration of two VLAs that could be used to construct two "horizontal virtual arrays," HVA1 and HVA2. We illustrate the use of the Turgut formalism to estimate the range ratio of an acoustic source from two vertical arrays operating in a dynamic littoral ocean waveguide. The data consist of broadband (60-140 Hz) acoustic signals that were projected to two VLAs during the SWARM95 New Jersey Shelf experiment. The source was 15.5 km from the VLA1 and 18 km from the VLA2, i.e., the range ratio is either 0.86 or 1.16. The source depth was 45 m, and water depths at the source location (VLA1 and VLA2) were 72, 70, and 90 m, respectively. Figure 5(a) shows the measured virtual array signal intensity along the VLA1 propagation track. Figure 5(d) shows the measured virtual array signal intensity along the VLA2 propagation track. In Fig. 5(a), the virtual array signal intensity was obtained by shifting the frequency bins of the received acoustic signals on the VLA2 up to ±11 Hz. Contours of constant virtual array outputs (level curves) were identified by an edge detection algorithm or by a visual marking, as shown in Figs. 5(a) and 5(d). Figures 5(c) and 5(f) show the cost function Φ defined as the rms value of a least-squares fit between the predicted level curves to the marked level curves. Minimum values of the cost function correspond to a best estimation of range ratios and the waveguide invariant parameter called beta. The best estimations were r2/r1 = 1.14, r1/r2 = 0.88, and β = 0.75. The estimated range ratios are comparable to the actual range ratios, r2/r1 = 1.16 and r1/r2 = 0.86 mentioned earlier. The agreement between the estimated and true range ratios is remarkable, especially considering that two VLAs have different propagation tracks and the environmental parameters are unknown along each track. Figures 5(b) and 5(e) show the levels curves calculated by using the estimated parameters and plotted on the measured virtual array intensity. Excellent agreement was obtained between measured and predicted level curves.

Figure 4 Image



FIGURE 4
Geometrical configuration and construction of two virtual arrays using two VLAs in a complex littoral environment.



Figure 5 Image
FIGURE 5
Illustration of curve fitting method for estimating range ratios and waveguide invariant β: (a,d) visual marking of the level curves; (b,e) comparison of measured virtual array output and calculated level curves by using estimated parameters; (c,f) cost function Φ as a function of range ratios and waveguide invariant β.

The impact of littoral fluid dynamic-induced waveguide variability on range ratio estimation has also been addressed. Figure 6(a) shows the SWARM95 experimental configuration and observed internal solitary wave (ISW) propagation direction. Figure 6(b) shows the virtual array output when the ISWs were outside the acoustic propagation track. Interference structures can be observed for almost the entire signal bandwidth and estimated parameters match the previously estimated values. Figures 6(c), 6(d), and 6(e) show the results when the ISWs were in the propagation track. Each figure is separated by 30 min. Interference structures seem to be destroyed by ISWs for signal frequencies greater than 100 Hz. Even in this case, a passive estimate of the range ratios was still possible for frequencies less than 100 Hz. Based on minor differences observed in the virtual array outputs, and estimated range ratios and β values, we conclude that the range ratio estimation technique is robust, even under highly dynamic oceanographic conditions.

Figure 6 Image
FIGURE 6
Virtual array outputs and calculated cost functions under highly dynamic oceanographic conditions. (a) Experimental configuration and observed ISW geometry; (b) ISWs are not present in the propagation tracks; (c,d,e) ISWs are in the propagation tracks.

Summary: We outline a robust method to passively estimate the range ratio of a target from two vertical line arrays in the littoral environment. We expect that the technique will allow reliable target range (and bearing) estimation if four vertical arrays are placed in an area of interest.

[Sponsored by ONR]

References
 1A. Al-kurd and R. Porter, "Performance Analysis of the Holographic Array Processing Algorithm, Ocean Acoustic Holography: Using a Reference Source to Remove Oceanographic Variability," J. Acoust. Soc. Am. 97(3), 1747-1763 (1995).
 2L.M. Brekhovskikh and Y.P. Lysanov, Fundamentals of Ocean Acoustics (New York, Springer, 1991), 2nd edition.
 3A. Turgut, "Method and Apparatus for Passive Acoustic Ranging in Shallow Water," Patent Application, Navy Case #84,558, 2003.