By Geoff Fein
The Office of Naval Research (ONR) is making investments to develop autonomous underwater mine hunting systems and to improve the ability of those systems to communicate through water, as well as remain on station longer, according to an ONR official.
“The leading edge issue is autonomy… the artificial intelligence that’s onboard [an unmanned underwater vehicle],” Tom Swean, organic mine countermeasures program manager at ONR, told reporters during a briefing earlier this week.
In an autonomous underwater vehicle (AUV) the sensors are like the eyes, ears and feeling. The brain affords the autonomy, Swean said. “We need to get that same sort of functionality in an underwater platform.”
AUVs need to act on data collected through sensors. In the mine hunting world the most efficient sensor is the sonar, he added.
“The AUV needs to act on that data. It needs to reason that it has detected the mine and needs to act on that data,” Swean said.
Communications and endurance are also important issues for AUVs, he added.
“For mine warfare, extraordinary endurance is not a KPP (key performance parameter). However, for something like distributed surveillance or long-term persistence for ASW (anti-submarine warfare), then, of course it is,” Swean said.
One area where endurance does play a role in the use of AUVs in mine warfare is power.
“If you want to have AUVs deployed by the Navy, then you need to pay attention to what type of batteries you have and whether or not they can be certified on various platforms,” Swean said. “Clearly that really limits us to more or less conventional battery technology, plus rechargeable lithium [batteries].”
Those types of battery technologies do allow several days operation, he added. “That’s enough for most things.”
Some of the things that are further down the road involve fuel cells…maritime fuel cell technology, Swean added. “They are in a more basic stage of development. It will be a few years before we can realize the fruits of that. There are some intense investigations going on.”
Another important aspect of AUV technology is the ability to get the systems to communicate. Swean noted ONR is at the point where they are developing AUVs for multiple vehicle systems.
“Sort of a cooperative autonomy. In other words, the UUVs are working together to accomplish a mission,” he said.
This means the systems need to communicate with each other, doing that generally in the water column, Swean said.
“There is a lot of work we have done in the past, and is ongoing, to develop underwater communications networks…acoustic networks…as well as higher data rates,” he said.
The issue is the bandwidth, Swean added.
“There is only so much bandwidth available in the water column. The acoustic velocities are nowhere near the speed of light. They are a small fraction of a small fraction of the speed of light,” he said. “So you don’t have the bandwidth and you have a huge latency and you need to accommodate…work that problem, with tactics as well as with further technology development.”
ONR will display many of its AUV systems, some developed for the Littoral Combat Ship, this month at AUVfest 2008, in Narragansett Bay, R.I.
Swean noted that ONR has been making strides in the areas of automatic autonomous target recognition, computer-aided detection and computer-aided classification.
“What we are doing is processing the sonar data in real time, and looking at the characteristics of each target that is detected,” he said.
Examples include looking at things like the shadow, which provides some shape information, looking at the dimensions, which provides length and width, and comparing that to conventional mine-like targets and making a call as to whether that target that has just been detected is mine-like, Swean explained.
“This is at range, so it’s not a positive identification,” he said. “That’s another phase of the mission we are working on where the cooperative autonomy, based upon these type of detections, either the vehicle itself will break off its current search path and go over and investigate the object more closely, or cue another vehicle, which has a higher resolution sensor to go over and look at it.”
Doing the onboard processing is very, very key to having an autonomous vehicle, Swean added. “You need to process on board.”
Trying to detect mines and differentiate them from other clutter scattered across the sea bed, is also challenging, Swean noted.
“From the point of view of detection and classification, there is an awful lot of clutter on the sea bed. There is a lot of man-made objects as well as a lot of rocks and things like that which can at times resemble mines,” he said. “So the biggest issue is separating the real object from the false alarms. We do put a lot of investment into higher resolution sonars, and into our processing, to try to be more clever in rejecting the false alarms because that is a big time consumer.”
The other issue that happens on the sea bed is that mines will tend to bury, Swean said.
“When they start to bury, even if they are partially buried, they have a much lower signature to a sonar,” he said. “So it has a lower signature and it’s harder to pull that signature out of the noise.”
These challenges have led ONR to explore other modes for detecting at range, Swean said.
For example, once a long-range sensor detects a mine-like object, that sensor or another sensor can be sent in to take a closer look at the object, and perhaps get close enough to use an electro-optical sensor or camera, to image the mine, he added.
“Even then, it may look like a mine but then again it may not be. We are not looking inside of the mine with any of these instruments,” Swean said.
Some of the greatest improvements have been in the area of sensors that can more clearly reject clutter and only accept a higher percentage of what are deemed real targets, Swean said.
“We have also worked our sensors so that they can penetrate the sea bottom, so we are now doing a much better job with buried mines,” he said. “Up until recently, the last three to four years, the only system that the Navy had that could detect a buried mine was a dolphin. We’ve now shown we can do that and, in fact, that’s one of the major type of systems that we are showing at AUVfest.”