1. Introduction
The underwater acoustic channel is a bounded medium with the sea surface and the sea bottom. In this bounded medium, the acoustic signals propagate through multiple paths including the direct path, surface reflected path, bottom reflected path, and other paths with multiple interactions from the surface and/or bottom. These different paths result in time spreading of the acoustic signal, which causes multipath interference. The acoustic signal is also affected by the Doppler induced by the motion of the source and/or receiver, surface variations, and water column variations. The multipath and Doppler affect the acoustic signal propagation and acoustic communications. This paper reports some water column effects that were observed during an experiment and studies their effects on the underwater sound propagation and underwater acoustic communications.
Underwater sound propagation is very sensitive to the water column properties like the sound speed profile (SSP). The SSP is a function of water temperature, salinity, and depth. It also changes with time and space, which causes time variability in sound propagation. These time variabilities can be due to many different physical phenomena happening in the water column such as internal waves, eddies, internal tides, or solitons. Internal waves are usually studied from synthetic aperture radar (SAR) images [
1]. Different image processing techniques have been applied on SAR images to study internal waves in a more effective way [
2]. Internal waves were observed in Norwegian waters and reported in [
3]. In [
3], these waves were characterized by a relative difference in the pixel value between crest and trough, wave width, and the number of crests. The typical pixel difference of 5 dB, wave width of 11 km, and four wave crests have been reported for Norwegian waters.
The effects of these water column variations on the underwater sound propagation have been studied in the literature. In [
4], an anomalous behavior was observed in the frequency response, which was caused by the interaction with the internal waves. In [
5], the SSP variations were reported in the middle of the water column, which was caused by the internal waves. In [
6], an internal wave bolus was detected in an SAR image. It was shown that internal waves introduced temperature variations in the water column, which resulted in the bending of the acoustic rays. The effects of water column variations on geo-acoustic inversion have also been studied [
7,
8,
9]. In [
7], shallow water sound propagation was modeled for a few kilometers in the presence of water column variations, and it was reported that the water column variations severely affected the estimated seabed properties. In [
8], a transform method was presented to model the effects of water column variations on the acoustic inversion. In [
9], the effects of internal waves on the geo-acoustic inversion were presented. It was shown that internal waves induced focusing and de-focusing of the sound intensity, which greatly affected the inversion results.
In this paper, the effects of water column variations on sound propagation and underwater acoustic communications are studied. In the literature, the physical phenomena in the water column have been studied using SAR images or thermistor chains. In this work, channel impulse response (IR) analysis is performed to show water column variations. In [
10], the effect of water column uncertainties on the non-coherent communication system was studied. The results showed that the system performance improved in the presence of the thermocline. The work presented in this paper is different from [
10] in a number of ways. Firstly, the time scale of these variations is different. The water column variations, observed in [
10], were on the scale of a few hours, while in this paper, the variations were observed on the scale of a few minutes. Secondly, the source-receiver positions and ranges were very different. In [
10], the source was deployed at the bottom, and the source-receiver range was 3 km. In this paper, the source and receiver were deployed in the middle of the water column, and the source-receiver ranges were 480 m and 670 m. Due to source-receiver positioning, different effects were observed and discussed.
In order to study the effects of water column variations on sound propagation, the channel IR estimates were studied. Two distinct features were observed in the arrival pattern: (1) the relative amplitudes of the first and second arrival were different at different receivers, and (2) strong amplitude variations were observed at the receivers due to the interaction between the acoustic signal and the water column variations. The Doppler analysis was also performed on the channel IR, which provided useful insight into the propagation of the water column variations. It was observed that a strong Doppler was induced due to the interaction between the receiver and the water column variations. The channel variability due to water column variations was presented by channel coherence analysis. The channel coherence along time at each receiver reduced due to time variability in the channel IR. The effect of water column variations on underwater communication was also demonstrated by a time reversal (TR) combiner. In the TR system, a known signal was transmitted for channel IR estimation. Assuming that all the multipath were cleared, the data signal was transmitted. The received data signal was correlated with the time-reversed version of the estimated channel IR. The performance of the TR combiner degraded in the presence of water column variations, which shows that these variations affect underwater communication and that there is a need for a robust technique to combat these variations. The main work of this paper is: (1) to study the effects of water column variations on underwater sound propagation using channel IR analysis, (2) to study the effects of these variations on underwater acoustic communications.
The paper is organized as follows: The description of the experiment is presented in
Section 2.
Section 3 describes the experimental data and results. It shows the channel IR analysis, channel coherence analysis, and the performance of the underwater communication system.
Section 4 concludes the paper.
2. Experiment Description
This section presents the description of the experiment, which was performed on 22 September 2016 in TrondheimFjord. During the experiment, four autonomous hydrophones were used.
Figure 1 shows the location of the source (Position B) and the receivers (Positions C and D) and the water depths at these positions. The source was deployed at a 20-m depth, and the water depth at the source position was 150 m. Two autonomous hydrophones were deployed at both Positions C and D in an array formation at 30 m and 60 m depth. The water depth was 200 m at C and 100 m at D. The source-receiver ranges at Positions C and D were 480 m and 670 m, respectively. Two sound speed profiles (SSPs) were obtained at different times at Positions B, C, and D and shown in
Figure 2.
The source signal was transmitted from the transmitter deployed on the NTNU research vessel R/V Gunnerus. Different sets of acoustic signals were transmitted during the experiment. The channel IR was estimated from the chirp signals, and the data signal was transmitted in the form of a binary phase shift keying (BPSK) modulated signal. In this paper, the signals between 14:11 and 14:20 were analyzed. The chirp signals were transmitted in the frequency band of 6320–10,680 Hz, and the data signal was transmitted in the 7000–10,000 Hz band. A total of 50 chirps were transmitted each minute where each chirp was 0.1 s long, and a silence of 0.2 s was placed between each chirp to clear the multipath. The data signal was transmitted for 40 s after the chirp signal transmission, and 5 s of silence was placed at the end of each minute. The same sequence was transmitted for 9 min.