DARPA's Sea Hunter to begin COLREGS testing in January 2017
Commencing in January 2017 DARPA will begin collecting data on how its Anti-Submarine Warfare (ASW) Continuous Trail Unmanned Vessel (ACTUV) conducts fully autonomous operations in preparation for international certification.Additionally, ASW track and trail testing of ACTUV could be delayed beyond fiscal year (FY) 2018.
ACTUV, now referred to as Sea Hunter, will begin the lengthy International Regulations for Preventing Collisions at Sea (COLREGS) certification process, an important step in determining whether the large unmanned surface vessel (USV) is a viable solution for conducting extended ASW missions.
COLREGS is a "threshold capability that is important for any large USV that will be operated at long distances from a human operator", explained Scott Littlefield, programme manager for DARPA's ACTUV programme.
"It is the rules of the road at sea," Littlefield told attendees at the annual AUVSI Unmanned Systems Defense maritime day, held on 24 October in Arlington, Virginia.
Sea Hunter is a 132-ft trimaran built by Leidos. The vessel was initially designed to autonomously track submarines for periods of weeks to months.
DARPA did not want ACTUV to be a remote-controlled vessel, particularly because of the reliability, latency and bandwidth issues associated with satellite communications, which would be used to control the vessel.
"You can't make that your achilles heel," Littlefield said. "You need something where if you lose communications with a remote operator it will continue to operate in a safe manner and complete the mission." For that capability, COLREGS is an enabler, he added.
COLREGS is also key to the concept of low manning because the platform can now go to Sparse Supervisory Control where one or two watchstanders could keep track of a fairly large number of Sea Hunters, Littlefield said.
"Because when they are out at sea in normal circumstances they are not required to have continuous interaction with an operator," he said.
Because COLREGS is such an important part of the Sea Hunter programme, DARPA didn't want to wait until they had a full scale prototype on the water to begin working on the capability, Littlefield noted.
DARPA developed algorithms for Sea Hunter autonomous operations and tested them in a system integration lab before very quickly taking them out to sea on a surrogate vessel, which has the same sensor, software, and computing plant as the full-scale Sea Hunter prototype.
"By doing that we were able to find out pretty quickly whether our COLREGS algorithms were working or not," Littlefield said. "[We are] at a point now that we feel comfortable that COLREGS is a solvable problem."
To achieve COLREGS there are a few things that have to be done, Littlefield noted.
Developers need to create a comprehensive world model of potential contacts, "because you can't avoid a contact you did not see", he said.
"It is not just about contacts but about classifying those contacts. For example, in COREGS, it matters whether the other vessel is a powered vessel or sailboat," Littlefield added. "So we are investing in some autonomous approaches for doing vessel classification and doing EO/IR [electro optical infrared] without having a human in the loop."
Another critical piece that applies to the long-term transition to this technology is developing a test and evaluation strategy.
Littlefield noted that Sea Hunter will never be able to conduct enough actual at-sea test hours to develop the statistical confidence in its ability to operate autonomously as effectively as a human operator.
"How much testing, how much evidence do you need [before you] say, OK it is good enough, we are ready to certify it is safe for operation," Littlefield said. "That is a hard problem. The way we will get to that is through some combination of modelling and simulation and actual at-sea testing."
Modelling and simulation can provide potentially millions of virtual hours at sea but researchers will still need to do the actual at-sea testing as part of the validation process.
"We are working through that right now. That will be a continuing challenge as this programme goes forward and becomes a navy programme," Littlefield said.
On a good day at sea DARPA can maybe set up four COLREGS scenarios, Littlefield noted. So in a week he estimates researchers can get 16 data points at a cost of USD10,000 per data point.
"So if you want to get statistics on hundreds of thousands of runs we can't get there from here with at-sea testing," Littlefield said.
DARPA has worked through a lot of basic testing on the full scale prototype.
Beginning in January, DARPA will begin collecting a lot more COLREGS data as it takes Sea Hunter out to sea for a week or two at a time, for almost every month through 2017, Littlefield said.
"We have a robust programme in place to continue to improve the COLREGS capability," he said.
FY 2017 is the last year of DARPA funding for Sea Hunter. Sometime during the fiscal year Littlefield will hand the keys to the US Navy's Office of Naval Research (ONR), which will take over the testing phase through FY 2018, a period that includes COLREGS, EO/IR sensors, and a few additional payloads, including for mine countermeasures (MCM) and anti-submarine warfare (ASW).
ASW testing for Sea Hunter will come under the auspices of ONR. Littlefield noted that testing may not complete by the end of FY 2018 due to schedule and funding issues.
"The flipside is, some of the payloads the navy is interested in are not ASW, so we have accelerated some of those and [are] putting ASW off until the navy gets around to it," he said. "Additional missions and payloads are already funded and there are a lot of other things we could do with this platform."While it will be up to the USN to decide whether or not to transition Sea Hunter into the fleet, there are some things Littlefield noted as important for the navy to get started on fairly soon.
For example, developing standards and policies for unmanned operations to provide a body of evidence to enable Naval Sea Systems Command or another command to declare that Sea Hunter is certified to go unmanned.
Another area is command and control (C2) of Sea Hunter. If the USV is launched and recovered from a vessel, the commanding officer of that ship would oversee the platform. But Sea Hunter will likely deploy more often from pier side; and while it might operate with a battlegroup it could also work independently. The navy will have to determine in those instances who Sea Hunter would report to and work for.
The only payload DARPA has demonstrated to date with Sea Hunter is a prototype of a low-cost, elevated sensor mast developed through the Towed Airborne Lift of Naval Systems (TALONS) research effort. TALONS is a towed parafoil designed to provide an ISR capability or communications relay payload at up to 1,500 ft above a ship. At that altitude, TALONS would extend a ship's view to about 40 miles.
news.usni.org
www.thedrive.com
www.defensenews.com
