SSCI and MIT Sea Grant have demonstrated AI that can take over senior ship officer roles on autonomous boats using the US Navy’s Unmanned Maritime Autonomy Architecture (UMAA). The AI was demonstrated on a fully autonomous Boston Whaler on the Charles River in Massachusetts.

The AI virtualises combined human shipboard functions of Lookouts, Sensor Operators, Quartermaster, Helmsman, and the Officer of the Deck. The software also combines computer vision, waypoint navigation, obstacle avoidance, vehicle control, and mission management.

MIT Sea Grant is a Federal-Institute partnership between the US Government and MIT. It runs two laboratories in Boston, including one devoted to autonomous underwater and surface craft. SSCI (Scientific Systems Company, Inc) is a Massachusetts based developer of AI for command and control functions and sensors.

The AI capabilities of the software are required for the Navy’s Unmanned Maritime Autonomy Architecture. This promotes “development of common, modular, and scalable software for Unmanned Maritime Vehicles (UMVs)”. It aims to reduce costs for acquisition programmes and speed up fielding of autonomous systems.

The software’s hierarchical levels of control open an evolutionary path for implementing increasing degrees of on-board intelligence for complex decision-making. During the demonstration, the autonomous Boston Whaler safely navigated the busy Charles River while piloting around another vessel using COLREG manoeuvres — International Regulations for Preventing Collisions at Sea — to avoid a collision.

MIT provided open-source Manoeuvre Operations software through MOOS-IvP (Mission Oriented Operating Suite – Interval Programming), performing local waypoint navigation, obstacle avoidance (COLREGs), and vehicle control. The AUV Lab at MIT Sea Grant provided the Boston Whaler.

SSCI provided the Mission Management software, which performed high-level reasoning, route planning, and executive oversight similar to human operator functions. SSCI also provided Perception and Situational Awareness software, which allowed the autonomous Boston Whaler to sense its surroundings with fused onboard sensors, including an electro-optical camera, LIDAR, and RADAR.

SSCI and MIT Sea Grant connected their autonomy software by packaging them in UMAA-compliant modules, allowing them to “speak the same language” without having to refactor the code to meet UMAA requirements. SSCI leveraged its UPSHOT software (UMAA’s Process SHOrtcutting Tool) to rapidly integrate the various components and make them UMAA-ready. UPSHOT supported system development, integration, and testing, allowing the team to field the capability quickly.