Autonomy Levels for Underwater Systems: A Taxonomy Proposal

A consortium of naval research labs has proposed a standardized autonomy taxonomy for underwater systems, addressing the inadequacy of SAE J3016 (designed for ground vehicles) for marine environments. The proposal covers UUVs, AUVs, and underwater infrastructure monitoring systems.

Underwater Autonomy Taxonomy: Why SAE J3016 Doesn't Work

SAE J3016's autonomy levels were designed for road vehicles in relatively predictable environments. Underwater systems present fundamentally different challenges that require a different taxonomy.

Communication constraints: Underwater acoustic communication has bandwidth measured in kilobits per second with latency measured in seconds or minutes. This fundamentally changes what "human in the loop" means — a level 2 equivalent for underwater systems may require onboard decision-making that would be level 4 equivalent on land.

Sensor diversity: Underwater sensing relies on sonar, acoustic Doppler, pressure sensors, and inertial navigation rather than cameras and LiDAR. The "operational design domain" concept needs corresponding redefinition.

Proposed taxonomy: The consortium proposes 6 levels from "remotely operated" (L0) through "fully autonomous with mission self-modification" (L5), with explicit treatment of the communication latency envelope at each level.

Standards path: The proposal is being submitted to ISO TC8 (ships and marine technology) for consideration as a new work item. Expect 2–3 years to formal publication if accepted.