Feature ArticleUNCLOS Under-Ice Survey: A Historic AUV Deployment in the High Arctic
By Tristan Crees
Chris D. Kaminski
Senior Systems Engineer
Port Coquitlam, Canada
In March and April, an International Submarine Engineering Ltd. (ISE) Explorer autonomous underwater vehicle (AUV), one of two built for Natural Resources Canada (NRCan), was deployed to the ice-capped seas in the Canadian high Arctic. Its mission was to conduct under-ice bathymetric surveys in support of Canada's submission to establish the outer limits of its continental shelf under Article 76 of the U.N. Convention on the Law of the Sea (UNCLOS). During this deployment, a record for under-ice endurance was set and several new technologies were demonstrated.
NRCan AUV at the main ice camp underwater docking station. (Photo courtesy of DRDC)
The NRCan AUVs are ISE Explorer-class vehicles with several innovative additions to make them suitable for arctic survey work. The baseline Explorer vehicle design is an evolution of AUV projects dating from ISE's Autonomous Remotely Controlled Submersible in the early 1980s and is commercial-off-the-shelf equipment. It is offered with a customization package to meet individual client requirements.
The most notable features on the NRCan AUVs include a 4,000-meter-depth-rated variable-ballast system, a 1,360-hertz long-range homing system and under-ice battery charging and data transfer capabilities. A short-range localization (SRL) system was also developed for close-range positioning. The homing and SRL systems were developed by Canadian defense scientists and engineers at Defence Research and Development Canada (DRDC).
The Explorer's range was extended from 300 kilometers to approximately 450 kilometers by adding an additional pressure hull section to accommodate extra batteries. The scientific payload on board included a Sea-Bird Electronics Inc. (Bellevue, Washington) SBE 49 FastCAT conductivity, temperature and depth sensor; a Knudsen Engineering Ltd. (Perth, Canada) single-beam echosounder; and a Kongsberg Maritime AS (Horten, Norway) Simrad EM2000 multibeam echosounder. In order to optimize battery endurance, the plan was to only turn on the EM2000 at strategic locations (i.e., potential sea mounts) along the mission path.
The expedition's task was comprised of a series of short and medium-range trials followed by three long-range missions to survey the seafloor. It was originally envisaged that more long-range dives would be undertaken, but poor weather delayed the deployment, significantly shortening the time available.
The launch-and-recovery site was based at a main camp near Borden Island (78 degrees 14 minutes north latitude and 112 degrees 39 minutes west longitude). One of the vehicles was deployed from an eight-by-three-meter hole cut through ice measuring between two and three meters. After completing short-range trials, culminating in a 100-kilometer test dive, the AUV began its transit to a remote camp 320 kilometers to the northwest. The AUV autonomously homed into the remote camp and was secured to a docking frame with the help of a small remotely operated vehicle. Without removing the AUV from the water, the batteries were charged and survey data was downloaded, all through a 1.3-by-two-meter hole in the ice.
Subsequently, a second survey mission was undertaken in a region known as the Sever Spur (about 79 degrees north latitude and 115 degrees west longitude) before returning to the remote camp. Finally, the AUV returned to the main camp for recovery.
In total, the AUV spent 12 days under the ice before being successfully recovered, and the AUV completed 1,000 kilometers of under-ice survey between the three missions. The AUV reached a maximum depth of 3,163 meters and transited at an average speed of 1.5 meters per second at an average altitude of 130 meters above the seabed.
Operating entirely under ice in such a challenging environment with the distances and objectives involved makes this a historic milestone in AUV development and polar science. ISE, DRDC, the Department of Fisheries and Oceans Canada and NRCan are now preparing for a 2011 deployment to collect additional Arctic survey data.
(Left) NRCan AUV at the docking station. (Right) ROV display of the AUV parked next to an ice keel at 79° N.
The NRCan Explorer AUVs were modified by ISE for Arctic operations. They feature a forward free-flooding payload section, a full-diameter pressure hull and a free-flooding aft section. The NRCan AUVs also have a variable ballast module located between the nose and payload sections. The vehicle control computer is located inside the pressure hull on a standard 19-inch PC rack. It is mounted on top of the vehicle battery trays. The NRCan AUVs were built to collect high-quality data that are of sufficient quality to form part of Canada's UNCLOS submission. Vehicle stability is an essential characteristic, and post-dive log data demonstrates a maximum deviation rate of 0.2° per second in the roll, pitch and yaw axes.
It was not possible to communicate with the vehicle during the under-ice missions. To enable recovery of the vehicle at a precise point and to overcome the problems created by navigational system error and ice movement during the mission, a homing system was developed.
The remote camp was located on an ice-floe drifting at up to 12 kilometers per day. Given the three-day duration of each mission, this meant the endpoint of the mission could be 36 kilometers away from the original planned location. Adding in the potential for error in the navigation system—particularly at high latitudes—meant a homing solution was needed that could function at ranges in excess of 50 kilometers.
The homing system consists of several elements. First, it features a custom seven-element hydrophone array designed by DRDC and manufactured under contract by Omnitech (Dartmouth, Canada). The array is installed in the nose of the vehicle, adjacent to the forward altimeter on vibration-isolating rubber mounts.
Second, software algorithms run on the ancillary homing and localization computer, mounted on the forward electronics tray inside the pressure hull. These algorithms take the raw hydrophone data and determine vertical and horizontal angles between the AUV and the detected sound source. These data are passed via an Ethernet link to the vehicle control computer, which implements a mission task verb to steer the vehicle toward the sound source. Custom task verbs were developed during the course of the project to enable the homing function to be included in the mission programming.
Finally, from the surface, the homing system also features a broadband acoustic projector (sound source), lowered through an ice hole, that emits a continuous tone at 1,367 hertz.
The end result is a system that worked well when enabled 50 kilometers away from the sound source, and it is thought to have a useful range greater than 100 kilometers in the under-ice Arctic environment.
One advantage of the system is that it generates accurate bearings even when the sound source is behind the vehicle, which means that if the AUV overshoots the sound source, it is able to turn around and stay in the vicinity. This is exactly what happened during the initial transit to the remote camp, and the system allowed for a successful recovery instead of needing to initiate a lost vehicle search process.
Charging and Data Transfer
To avoid needing to recover the AUV at the remote camp location, an underwater battery charging system was employed. Once the AUV had been secured to the underwater dock, a SEA CON® (El Cajon, California) CM2012 Series underwater mateable connecter was lowered and manually attached to the mating point. This provided a 100 Base-T Ethernet connection for data transfer and mission planning procedures. The connector also contained the battery-charging mechanism, which required 12 to 15 hours to charge a fully depleted battery.
In general terms, mission planning (MP) means all of the activities that must be planned and taken into account for a successful AUV mission.
Each mission plan was created with the route data, depth set points and commands for payload sensors and contains all of the commands to complete the mission objectives without operator intervention.
The MP file also contains a plan to cater to emergencies and system faults, enabling the AUV to respond accordingly to changes in vehicle status throughout the dive.
Electronic-chart-based software is essential to efficient and safe MP. The Mimosa package, developed by Ifremer specifically for AUV MP, was used. Mimosa allows geographic information system layers, such as bathymetry and/or satellite photos, to be correlated with chart data and used to graphically draw the desired AUV path, which is then exported as an MP file.
Before the mission plan was downloaded to the vehicle, it was subject to a formal review process that required at least two qualified people to examine the file in detail.
The final approval to commence a mission was always the responsibility of the senior technical authority (DRDC) representative on site. The team developed a mission-verification checklist that included confirming each position along with all of the survey modes, contingencies and the total mission length. Both reviewers signed and dated a copy of the checklist for each new mission. While this verification process does not guarantee that the mission will be perfect, it does cut down on the number of operator errors.
Risk Assessment and Mitigation
While it was obvious that the risk in completing a series of under-ice missions was significant, quantifying that risk was difficult. The challenge was to identify the individual risks and to develop mitigation that would optimize the probability of successful mission completion.
During the preparation for the 2010 deployment, the U.K. National Oceanographic Centre (NOC) convened a risk assessment workshop in Halifax, Canada. Under the guidance of the NOC's Gwyn Griffiths, participants from NOC, Ifremer, the University of Southern Mississippi, DRDC and ISE met to discuss failure modes and the likelihood of recovery following an emergency. The data was processed using the Kaplan-Meier method to produce a probability of success that was used in the mission planning phase to help reduce risk while preserving productivity. One of the key findings of the workshop was that if an AUV operates properly at depth for more than two hours at the start of the dive, it is likely to continue doing so for the remainder of the mission.
To prove the enhanced endurance capabilities of the AUV, a series of 60 and 72-hour test dives were conducted prior to the deployment.
The Explorer is designed with many safety systems, including a comprehensive fault management and response system designed to protect the vehicle in the event of equipment failure. The vehicle is also designed with modular subsystems such that the failure of any one component does not necessarily cause the failure of the vehicle as a whole. Furthermore, the control software uses a hybrid-hierarchical scheme that allows some actions to preempt others. For example, collision avoidance takes priority over altitude keeping.
Results and Performance
In all, 1,000 kilometers of sonar data were collected under the ice during three missions, all of which were conducted without the need to recover the vehicle between missions. During the 240-hour operation, the AUV ran well with only minor faults that it was able to manage without intervention. The missions were successful due to the robust design as well as the meticulous planning and preparation undertaken by the joint project team.
Bad weather had the largest effect on operations. Several days were lost due to the inability to transport equipment to the camps, an unavoidable aspect of operating in harsh conditions.
The success of this deployment demonstrates that AUVs are ready for long-range, unsupervised or unescorted missions in harsh environments. This opens up a realm of potential applications, including oil-flow mapping. With regard to recent events in the Gulf of Mexico, this may be of particular interest in the near future. However, the next deployment for the two AUVs is a planned return to the Arctic in 2011 for further under-ice UNCLOS missions.
Jean Marc Laframboise and Richard Mills of ISE also contributed to this article, along with Alexander Forrest of the Department of Civil Engineering of The University of British Columbia, Jeff Williams of the University of Southern Mississippi, and Erin MacNeil, David Hopkin and Richard Pederson of DRDC.
The authors thank NRCan's Earth Sciences Sector, DRDC Atlantic and the Polar Continental Shelf Project for their roles in this work.
Tristan Crees graduated with honors from Monash University in Australia with a bachelor's degree in computer science and engineering in 1996. He joined International Submarine Engineering Ltd. in April 1998 as a robotics engineer, and he was a key player in the design of the Explorer autonomous underwater vehicle (AUV) software and control system. He was the project manager for the construction of the Explorer AUVs for NRCan.
Chris D. Kaminski joined International Submarine Engineering Ltd. in 1991 after completing an electrical engineering degree at The University of British Columbia. He has been involved in many autonomous underwater vehicle (AUV) projects and deployments, including the Theseus under-ice cable laying mission in the Canadian Arctic in 1996. He managed the NRCan AUV arctic upgrades and 2010 Arctic deployment and is currently the project manager for the 2011 NRCan Arctic deployments.
James Ferguson has been responsible for the overall operation of the company's autonomous underwater vehicle development since 1981. He has also served as an adviser to the Minister of National Defence, a director of the Canadian Manufacturing Association and the chairman of the British Columbia Technology Industry Association. Currently, he is the president of the Engineering Committee on Oceanic Resources.