On August 10, 2024, about 1912 local time, a National Oceanic and Atmospheric Administration (NOAA) Great Lakes Environmental Research Lab (GLERL) SeaTrac SP-48 uncrewed surface vessel (USV) and a privately owned 45-foot-long Prestige 450 Flybridge recreational vessel collided while underway in Lake Erie, 6 miles east-southeast of Monroe, Michigan. The USV became fouled in the recreational vessel’s starboard propeller pod, and the two vessels were towed conjoined to a nearby marina and later separated. No pollution was reported. One minor injury was reported. The USV was considered a total constructive loss valued at an estimated $186,000. Damage to the recreational vessel was estimated at $100,000.

​GLERL operated the SeaTrac SP-48 USV in an area offshore frequented by recreational vessels. The vessel was about the size of a small recreational vessel (15.7 feet long) and designed with only about a foot of its hull visible once in the water. Given these conditions, it would have been difficult for operators of other vessels to perceive the USV. Additionally, the USV likely could not be perceived by recreational vessels via electronic means, such as radar or an automatic identification system, nor did it have auditory means to signal its presence to other vessels. We found that, given the USV’s area of operation, the vessel’s visual, electronic, and auditory perceptibility were inadequate. Had GLERL tested the USV’s perceptibility to operators of other vessels typically in the USV’s area of operation, GLERL could have identified the USV’s inadequate perceptibility and determined that the USV was a navigational hazard for surrounding vessel traffic. ​

At the time of the collision, there were no industry-wide perceptibility standards specific to USVs operating in US inland and territorial waters. We found that additional standards are necessary to enhance and ensure the perceptibility, collision avoidance capability, and navigational safety of USVs.

GLERL operated the USV in “adapted hold” mode, wherein the USV would automatically move back to set waypoints if it drifted outside a 150 foot radius set by the pilot. When the vessel was in adapted hold mode, the pilot passively monitored the vessel. We found that this configuration and operation of the USV without active monitoring left the USV without a proper lookout. Combined with the USV’s inadequate perceptibility, the lack of proper lookout rendered the USV vulnerable to collision in the area where it was operating. 

When the USV was in adapted hold mode, GLERL passively monitored the vessel, relying on the USV to automatically maintain the waypoints set by the pilot. The USV could not calculate decisions regarding navigation or operational action and thus could not take appropriate action to avoid a collision. We found that, given the USV’s low level of autonomy in adapted hold mode, GLERL’s monitoring of the USV was not sufficient to safely operate it in an area offshore frequented by recreational vessels.

We determined that the probable cause of the collision between the NOAA GLERL USV and recreational vessel was the inadequate perceptibility of the USV to nearby vessels, given the operating area and conditions, and the lack of active monitoring given the vessel’s configuration, mode of operation, and level of autonomy, which left the USV without a proper lookout to avoid a collision.

​​As a result of this investigation, we make the following new safety recommendations.

To the US Coast Guard:

• Develop and promulgate a national framework of perceptibility, collision avoidance capability, and navigational safety standards for uncrewed surface vessels. (M-26-03)
• Incorporate perceptibility, collision avoidance capability, and navigational safety standards for uncrewed surface vessels into regulations. (M-26-04)

To the National Oceanic and Atmospheric Administration:

• Revise your “Uncrewed Marine Systems Manual” to ensure all uncrewed surface vessels are perceptible to vessel traffic given the vessel’s area of operation. (M-26-05)
• Incorporate into your “Uncrewed Marine Systems Manual” a requirement for testing of uncrewed surface vessels you operate to ensure adequate perceptibility to operators of other vessels typically in the uncrewed surface vessel’s area of operation. (M-26-06)
• Require your uncrewed surface vessels to be monitored at a level commensurate with the uncrewed surface vessel’s actual level of autonomy, factoring the vessel’s area and mode of operation. (M 26 07)
  

​To SeaTrac Systems, Inc.:

• ​Modify the design of the SP-48 uncrewed surface vessel to further enhance its perceptibility as practicable given the nature of the uncrewed surface vessel’s work and area of operation. (M-26-08)