A recent month-long Coast Guard evaluation of commercially available unmanned surface vehicles (USVs) off the Coast of Hawaii demonstrated the effectiveness of the technology as a force multiplier in helping to carry out maritime domain awareness (MDA) missions in remote ocean areas but further advances are necessary, the Coast Guard said on Monday.
The initial results from the evaluation showed that the USVs can provide “some level of daytime MDA” over an extended period but onboard artificial intelligence (AI) and machine learning capabilities are necessary to speed data processing and limit data transfer to actionable information, the service said. The service also said that limiting data transfer to actionable information will save on the use of expensive bandwidth.
The types of missions that USVs with different sensors could perform in remote areas include combatting illegal, unreported and unregulated fishing, search and rescue, autonomous search patterns, surveillance and reconnaissance, screening vessel traffic, and a full-range of law enforcement operations.
The Coast Guard in early November concluded its evaluation of two commercially available USVs, Saildrone’s wind-powered Saildrone and Spatial Integration Systems’ diesel/solar-powered Watcher cutter boat, which both had strengths and weaknesses.
“This evaluation showed that using autonomous USVs for future persistent MDA efforts will likely require a layered solution,” Scot Tripp, project manager with the Coast Guard’s Research and Development Center, said in a statement. “USVs like Saildrone are capable of performing MDA missions for up to a year without maintenance, but their low transit speed does not allow them to pursue a target of opportunity to collect more information. In contrast, the Watcher is only capable of 30 days endurance but has the capability of traveling at speeds over 30 knots. It can also be instructed to pursue a target upon detection for better imagery. A system where these USVs worked together could prove to be a valuable tool for future MDA capabilities.”
The USVs proved they can detect vessels more than a mile away and in certain situations beyond four miles.
AI Use
The Saildrone was equipped with AI software whereas the Watcher was not.
“The camera array captures imagery every five seconds and completes onboard analysis of each image using Saildrone’s Artificial Intelligence software which has been trained on vessel detection,” says an initial Coast Guard “Quick Look” analysis of the evaluation. “Upon the AI determining a vessel has been detected, the drone sends the image, line of bearing to the detected vessel, and its own location to the Saildrone command centers. To conserve satellite bandwidth and ensure the command center is not inundated with alerts, the AI is configured to send only the clearest images approximately every 15-20 minutes. All other images are saved onboard and can be later accessed.”
For the Watcher, “The system images every 10 seconds, however it did not possess any AI capability to determine which images should be sent with the alerts and instead would send imager as it was captures,” the Quick Look says. “To conserve bandwidth, the satellite transmission of imagery from The Watcher was intermittently turned on just to demonstrate the vessel’s capability.”
The Coast Guard used contractor-owned and operated business models with the Saildrone and Watcher vessels. The Research and Development Center also added its own USV to the mix, a 29-foot craft based on its Response Boat-Small II.
Government-Owned and Operated
The 29RDC was operated by junior enlisted boat drivers to senior officers who interactively planned and executed their own missions in real-time. They Coast Guard USV was also operated by RDC watchstanders in New London, Conn., demonstrating control of the vessel from 5,000 miles away using cellular service.
The Response Boat-Small is provided by Metal Shark and was outfitted with the SM300 vessel intelligence system from Sea Machines to enable operator-in-the-loop autonomous command and control. The 29RDC was controlled by laptop using the Sea Machines’ user interface and the craft was configured with cameras off the bow, stern and withing the cabin for situational awareness.
The demonstration was conducted about the same time the National Academies of Sciences issued a report saying the Coast Guard needs to be more aggressive in examining and adopting unmanned systems to help fulfill its missions. Coast Guard Commandant Adm. Karly Schultz, based on the report, has directed his operational directorate to lean-in in on potential uses of unmanned systems technologies.
The service is currently equipping all of its National Security Cutters with drones that can remain aloft for 12 hours at a time, helping the ships extend their surveillance capabilities and save the use of onboard helicopters for when they are absolutely needed.
The results of the USV study will also help inform policy decisions related to the increasing use of the vessels in the marine environment.