status, FAU dissertation, M Kobold

Current status (submission of "Acoustic Error Approximation due to Gouy Phase in the sea" #ADV23-AR-01652-TR:

Submitted to AIP Advances and Login, awaiting decisions from reviewers and AIP-Adv.

Conference paper presented October 2022:

IEEE OCEANS 22 Abstract and download , The paper was presented in October 2022: Kobold, Michael C., and Pierre-Philippe Beaujean. "Background Structure Functions for statistical acoustic propagation characterization." In OCEANS 2022, Hampton Roads, pp. 1-10. IEEE, 2022.

USA (USPTO) Patents related to this dissertation are
US11431421 (30au22)
and
US11653125 (16my23).

Using the data measured on 05 August 2007, one of the structure functions, distortion D(z1, z2), appears as a surface for two depths at the end of spatial periods as published for the entire day, for the structure function, The effect from the spread of full day set of SSPs is apparent. Please see the IEEE OCEANS 22 conference paper for the dissertation for explanations of distortion metrics; they are far beyond the scope of a web page. The SSPs show some downwelling, especially from the surface, hence the label for this set is "the bottom channel." However, the SSPs for this 05aug07 data are clustered in two sound speed groups. Further, when the groups are plotted independently, as follows, sub-clusters appear. This can happen when the day starts cloudy (cold water, low sound speed) but then the sky clears, heating up the canals and shipping channels in the Dania Beach marina and the Port Everglades facilities.
For the slower speed "bottom" channel SSPs The slow-speed restricted 05aug07 structure function has less distortion since D_{max}(z_1, z_2) \approx 2.5E-6, a dimensionless metric, normalized to reference sound speed 1500 m/s, than the D_{max} ~ 7.6E-6 calculated using the entire day's SSP collections shown in the prior paragraph.
The "high-speed" structure function shows a higher distortion of D_{max} ~ 4.2E-6 for the wider spread of speeds in the high-speed cluster of SSPs that appear to come from measurements later in the day.

A detail: propagation studies via ham radio programs

Low priority: Programs such as WSJT-X control and log short conversations in several different modes, modulations, protocols, and frequencies in order to detail propagation statistics for various regions. Since I must use them for work, I will see if they can be used for UW aComms to automate propagation measurements.

Another detail from the status:

The bandwidth W could be the inverse of our signal preface carrier at 16 kHz. For a 5 kHz bandwidth (Benthos C band) this provides TW = 3.2. To get a TW of unity we could reduce the W or increase the time, tau, such that W = 0.06 ms x 17 kHz. The symbol time-bandwidth of TW = 1 is used in the paper [b] for qualitative behavioral/sensitivity characterizations. A TW bandwidth sketch provides a TW bandwidth coordinate system in order to provide a foundation for the investigation. A signal has a time evolution downward on the waterfall charts. This is usually an evolution in time. However, in order to see how different frequencies propagate with different Doppler frequency results, plot the signal spectrum as a function of propagation distance, x. From the results at the receiver (RCVR) we can multiply a mean square Doppler spread, B, by the Delay, L, to get BL products for different times or locations.
The result is a qualitative understanding of the effect on channel capacity , C.

These Notional sketches below show that TW product a is less than TW product b. The waterfall style time flows downward on the ordinates. The frequency increases to the right on the abscissae. However, in this restricted case a scale of 2 acted only on the height of the spectrum, doubling for b the height of the same shape a. Therefore, in this figure, BLb = 2 BLa. In reality the shape of subsequent received signals will not be exactly the same shape, a scaled version thereof, and in some cases not even closely similar:

The paper [b] provides a qualitative view of the effect of time-bandwidth product TW along with the Doppler-delay product, BL, on the channel capacity, C. The following figure is a qualitative notional sketch of how the BL product affects C when the TW is approximately unity for the assumptions contained in the paper[b].

The following figure sets up one current investigation into the Doppler frequency shift and time delay variation of the received spectrum with respect to the propagation paths and time of arrival. The frequency is on the abscissa, as with the first two figures. However, the ordinate, still pointing down as with waterfall charts, is the range, x, along the propagation trajectory. These are variations that act on the same signal, ideally an impulse, in which case we could have a useful impulse response in tau time, h(f, tau).

Background Structure Functions for Underwater Acoustics to forecast Phase Statistics and Reduce Energy Loss, starting with investigation of paper [b]: Presented to FAU Ocean and Mechanical Engineering Department interim chair Pierre-Philippe Beaujean, PhD, Professor. for Michael C. Kobold

[a] Proakis, John G., and Masoud Salehi. Digital communications. Vol. 4. New York: McGraw-hill, 2001. [Is this Vol. 4 or Version 4, the latter which I checked out from the NSWC PCD tech lib. Found my personal Version 2 that the library excessed into my collection.]

[b] Beaujean, Pierre-Philippe J., and Lester R. LeBlanc. "Adaptive array processing for high-speed acoustic communication in shallow water." IEEE Journal of Oceanic Engineering 29, no. 3 (2004): 807-823.

[c] Harry L. van Trees, "Radar-sonar processing and Gaussian signals in noise, Part III: Detection, Estimation, and Modulation Theory," Wiley-Interscience, a John Wiley & Sons, Inc. Publication, New York,2001.

[d] Harry L. van Trees "Optimum Array Processing, Part IV: Detection, Estimation, and Modulation Theory," Wiley-Interscience, a John Wiley & Sons, Inc. Publication, New York, 2002.

Status of Phase Characterization for Underwater Acoustics