Every ship that drags a listening cable behind it works from a quiet assumption: that the cable runs straight. The sea rarely cooperates, and currents and drag bend the line until the directions it reports begin to drift. In a new open-water trial, a team at Harbin Engineering University has built a cable that sensed its own shape using the same optical fibre that carries its sound, so the instrument now knows where its parts truly are while it listens. It is a small correction, but it reflects a broader habit of engineering: letting a tool measure its own distortions instead of pretending the
A Towed Hydrophone Cable That Measures Its Own Bend Improves Sound Bearings and Wave Readings at Sea
A towed array that calibrates itself in place
Mark
So what's actually new here? Towed arrays already exist.
Mimi
The new part is that the cable knows its own shape. Normally a few depth sensors sit along it and the rest is guessed. Here the fibre itself reads depth along 160 m.
Mark
Why does the bend matter?
Mimi
Beamforming assumes element positions. If the cable bows, the focus smears and the bearing is biased. Using the measured depth profile raised peak power by up to 4.10 dB.
Luke
Up to 4.10 dB, though. That's the best case, on 1 kHz pulses from a source 5.7 km away. I wouldn't read it as a typical gain.
Mimi
Fair. And the bearings were checked against a GPS reference in two bands, vessel noise at 120 to 150 Hz and those pulses.
Mark
The wave readings seem like a side benefit.
Mimi
A useful one. Wave heights of 1.41 to 2.66 m landed within 4.31 to 11.03 percent of the Copernicus model, from the same low-frequency signal.
Luke
Note the comparison is to a model, not to a wave buoy. That's a looser yardstick. And it was one vessel, one speed, one sea state.
Mark
So what's unknown?
Luke
Shallow and coastal water, heavy weather, other tow depths. The authors say accuracy may drift there, and nothing in the trial settles it.
Der Puls
- A towed hydrophone cable, 770 metres long, is trailed behind a ship while its software assumes a straight line that the sea keeps bending.
- Bent geometry smears the focus of the array, so the bearings it reports to a listener can quietly be wrong.
- The fibre itself now doubles as a depth gauge, its readings matching a pressure sensor with an R² of 0.99 and revealing the cable's true bow.
- Feeding the measured shape back into the model lifts peak beamforming power by up to 4.10 dB, though trials covered only one vessel, speed and sea state.
Every ship that drags a listening cable behind it works from a quiet assumption: that the cable runs straight. The sea rarely cooperates, and currents and drag bend the line until the directions it reports begin to drift. In a new open-water trial, a team at Harbin Engineering University has built a cable that sensed its own shape using the same optical fibre that carries its sound, so the instrument now knows where its parts truly are while it listens. It is a small correction, but it reflects a broader habit of engineering: letting a tool measure its own distortions instead of pretending they are absent.
A ship towing a long hydrophone cable has a blind spot. Currents and drag bend the cable, yet the software that converts its signals into directions treats it as a straight line. Researchers at Harbin Engineering University have now built a towed array that measures its own depth and curvature while it listens, drawing both jobs from the same optical fibre.
The old tension is simple. A long array sharpens bearings, and towed arrays are prized for being passive, cheap and easy to launch. But beamforming needs to know where each element really sits. Conventional arrays carry only a few depth sensors, so the shape between them is guessed, and a bent array blurs its focus and biases its bearing.
The new 770-metre cable splits its work between two sections. One does the sensitive listening; the other reads optical phase along the fibre and, because water pressure shifts that phase, reports depth. Against a pressure sensor, the depth readings agreed closely, and the reconstructed profile showed the cable bowing rather than lying flat. Wave measurements landed within a few percent of a standard wave model, and when the measured shape was fed back into the array model, beamforming power rose by up to 4.10 dB.
The limits are candid. The trials used one vessel, one speed and one sea state, and coastal and shallow water remain untested. The next question is whether a cable that calibrates itself in place can keep doing so across more ships and rougher seas.
A ship towing a long hydrophone cable has a blind spot. Current and drag bend the cable, but the software that turns its signals into bearings treats it as a straight line. A team at Harbin Engineering University in China has now built a towed array that measures its own depth and curvature while it listens, using the same optical fibre that carries the sound.
The problem is old. Towed arrays are valued because a long aperture sharpens bearing estimates, and because they are passive, cheap and easy to launch. Beamforming is the step that converts arrival delays across the elements into a direction, and it needs to know where each element really is. Conventional arrays carry only a few depth sensors, so the shape between them is guessed. A curved array then smears the focus and biases the bearing.
The new cable is 770 m long and is towed about 500 m astern. It splits the work between two sections. An interferometric section of 128 elements spans 104 m, each element wound with 22 m of fibre, and does the sensitive listening. A distributed acoustic sensing (DAS) section has 32 elements at 5 m spacing over 160 m. It reads optical phase along the fibre and doubles as a depth gauge, because hydrostatic pressure shifts that phase. Both sections sit on the same air-backed mandrel, 42 mm long and 11.6 mm across. The data are separated by frequency: below 0.02 Hz for depth and shape, 0.02 to 0.3 Hz for waves, and 5 Hz to 2 kHz for sound.
The depth readings held up. Against a pressure sensor, DAS-inferred depth gave a coefficient of determination of 0.99 and an average deviation of 0.80 percent. The reconstructed profile showed the cable bowing rather than lying flat. In the wave band, the measured spectral peaks were 0.102 and 0.098 Hz. These sat near the expected 0.105 and 0.107 Hz once towing speed and heading were accounted for, differing by 2.86 and 8.41 percent. Significant wave height, retrieved from the same spectra and corrected for element depth, ranged from 1.41 to 2.66 m, within 4.31 and 11.03 percent of the Copernicus wave model. At 50 m tow depth, the authors estimate a detection floor of 0.15 m and a dynamic range of 180 dB.
For sound, the DAS section ran 11.57 dB above its own land-quiet baseline across 5 to 2,000 Hz, while the interferometric elements were quieter above about 2 kHz. Beamforming tracked a GPS reference bearing in two bands: vessel noise at 120 to 150 Hz, and 1 kHz pulses from a source 5.7 km away. When the measured depth profile was fed into the array model, peak beamforming power rose by up to 4.10 dB.
The limits are plain. The trials were open-water deployments in December 2025 and January 2026, with one vessel, one speed and one sea state. Coastal and shallow water have not been tested. Depth and wave retrieval depend on pressure, so accuracy will likely drift with tow depth and heavy weather.
What the trial does show is a towed array that calibrates itself in place. The authors point to wave monitoring from vessels of opportunity, fisheries and marine-mammal acoustics, and any sonar that must work without a straight line. The next test is whether the results hold across more ships, speeds and rougher seas.
Bemerkenswerte Zitate
Depth and wave retrieval lean on pressure, so accuracy will drift with tow depth and heavy weather.— Study authors, as reported by EurekAlert!
Möchten Sie die ganze Geschichte?
Das Original lesen bei EurekAlert! ↗