Platform & Operasi Survei
Under-Ice Survey: Mapping the Seabed a Vehicle Cannot Surface to Check
Every other autonomous underwater vehicle survey in this series has one safety net: if navigation error accumulates too far, the vehicle can surface, get a GPS fix, and correct itself. Under ice, that option does not exist. There is no clear patch of open water waiting overhead — only a solid ceiling that the vehicle cannot break through and, if its dead-reckoning drifts far enough off course, may not even be able to find its way back to the one hole it launched from. Under-ice survey is autonomous underwater vehicle work with the single most-used failure recovery mechanism deleted, and the entire discipline is built around compensating for that fact.
The Problem Every Other AUV Survey Doesn't Have
In open water, an AUV's dead-reckoning position estimate — built from heading, speed, and elapsed time — drifts steadily worse the longer it goes uncorrected, which is exactly why AUVs surface periodically for a satellite fix. In a GPS-denied, ice-covered environment, that correction is unavailable for the length of the entire mission underneath the ice sheet. Accumulating navigation error in that setting doesn't just degrade data quality; without a reliable acoustic position fix from an external reference, the vehicle can become, in the words of researchers who work on the problem, "increasingly lost" — a genuinely different risk category from a wandering GPS track in open water.
Where It Started: A Vehicle Lowered Through a Hole in the Ice
The earliest purpose-built under-ice vehicle was the Unmanned Arctic Research Submersible (UARS), developed by the University of Washington's Applied Physics Laboratory and deployed beneath Fletcher's Ice Island near the North Pole starting in spring 1972, after preliminary testing in Lake Washington. UARS was rated to 450 m depth with 12 hours of endurance, weighed 410 kg, and had to be lowered through a 4 m hole cut in the sea ice using a gantry. Its navigation and communication system was a 50 kHz long-baseline (LBL) acoustic array made of four free-floating transducers deployed through the ice near the launch site — the direct ancestor of the acoustic positioning networks still used under ice today.
Modern under-ice navigation still leans on the same core idea — acoustic ranging from fixed or semi-fixed reference points — but layers far more onto it. Long-baseline systems now commonly pair with a Doppler velocity log for bottom-track speed and an inertial navigation system for dead-reckoning between acoustic fixes, so that position error grows with distance traveled rather than compounding purely with elapsed time. More recent systems replace fixed seafloor transponders with buoys frozen into or drilled through the ice, carrying acoustic modems that trilaterate the vehicle's position from one-way travel times, with the model correcting for the actual sound-speed profile of the water column in real time rather than assuming it is constant.
Case Study: Twelve Days, a Thousand Kilometers, and a Legal Claim
In March and April 2010, an International Submarine Engineering Explorer-class AUV, operated for Natural Resources Canada, spent 12 consecutive days under Arctic ice, surveying close to 1,000 km of track before recovery — a record-setting deployment at the time. The mission's purpose was not purely scientific: the bathymetric data supported Canada's submission to extend its continental shelf claim under the UN Convention on the Law of the Sea, a process that depends on precisely the kind of seabed depth and shape data that satellites and surface ships cannot collect beneath permanent ice cover. The 2010 deployment built directly on earlier Canadian work: from 1993 to 1996, the Theseus AUV, developed under the Defence Research and Development Canada "Spinnaker" program, had already completed two 200 km under-ice missions launched from CFS Alert on Ellesmere Island, proving that long-range, GPS-denied missions under Arctic ice were achievable before the sensors and navigation systems existed to do it routinely.
Case Study: Sixty Kilometers Into an Ice Shelf Cavity
In January 2009, the British Antarctic Survey and National Oceanography Centre Southampton sent the Autosub3 AUV — a 7 m, 3.5 tonne vehicle rated to 1,600 m depth and powered by roughly 5,000 D-cell batteries — beneath the floating ice shelf of Pine Island Glacier in West Antarctica. Over six missions totalling more than 500 km of track and roughly 94 hours of runtime, Autosub3 penetrated up to 60 km into the cavity beneath the ice, well past where any ship or satellite could reach, using onboard multibeam sonar to map the underside of the ice shelf and the seabed beneath it simultaneously. The survey revealed a previously unknown ridge on the seabed roughly midway between the ice front and the grounding line, its crest standing about 700 m below sea level — a feature now understood to have influenced the glacier's retreat history. The mission produced what researchers described at the time as roughly 90% of all data ever collected by AUVs beneath ice shelves, from a cavity a ship could never have entered.
A Discipline Defined by What It Can't Do
Fifty years after UARS was lowered through a hole in the ice near the North Pole, the core constraint hasn't changed: a vehicle under ice has to know where it is without ever being able to check against a satellite, and it has to get itself back to a known exit point without that same check. What has changed is how much redundancy engineers now build around that single constraint — acoustic networks, inertial aiding, Doppler bottom-tracking, and real-time sound-speed correction, stacked together specifically because no one of them, on its own, is trusted to carry a mission across 60 km of ice-covered ocean and back.
References
- British Antarctic Survey, "Robot Submarine Searches for Signs of Melting Under Antarctic Ice Shelf," https://www.bas.ac.uk/news/robot-submarine-searches-for-signs-of-melting-under-antarctic-ice-shelf/
- "12 Days Under Ice – An Historic AUV Deployment in the Canadian High Arctic," IEEE/OES Autonomous Underwater Vehicles Conference, 2010, https://ieeexplore.ieee.org/document/5779651/
- Randeni, S.A. et al., "A High-Resolution AUV Navigation Framework with Integrated Communication and Tracking for Under-Ice Deployments," Journal of Field Robotics, 2023, https://onlinelibrary.wiley.com/doi/full/10.1002/rob.22133
- Jakuba, M.V. et al., "Long-Baseline Acoustic Navigation for Under-Ice Autonomous Underwater Vehicle Operations," MIT, https://web.mit.edu/2.166/www/handouts/jakuba_fsr08_submission.pdf
- "Unmanned Underwater Vehicles in Arctic Operations," ResearchGate, https://www.researchgate.net/publication/330703330_Unmanned_underwater_vehicles_in_Arctic_operations
Related Articles
Autonomous and Unmanned Survey Platforms: Taking the Crew Off the Boat
AUV Mission Planning: Budgeting Battery, Range, and Communication Before the Vehicle Ever Leaves the Deck
