SBIR-STTR Award

The US Navy has embarked on an effort to diversify its fleet to include fully unmanned surface platforms. As the endurance of these unmanned systems extends from days to months, pre-mission planning will be insufficient to handle uncertainty in operations. Nowhere is the resultant need for adaptive planning more pressing than in weather avoidance, where the dynamics of nature limits practical forecasting to no more than ten days of validity. To account for changing and uncertain conditions, advanced planning and re-planning methods need to be developed to allow for robust and safe extended duration operations that incorporates machine forethought, where the platform avoids entering situations that dangerously restrict possible future actions. For example, while following the edge of a reef may appear to be the shortest path around it, a mis-prediction as to the seakeeping ability of the platform in the face of adverse winds could be devastating.

Benefit:
Metron's proposed algorithm enhances the ability of autonomous platforms to operate for extended duration missions and provides (1) reduced risk to platform, (2) reduced oversight necessary from human operators, and (3) improved perception of balancing risk with mission success,. The proposed system has natural commercial extensions to improve safety, reliability, and efficiency in the areas of scientific research, geophysical surveys, oil and gas, and commercial shipping.

Keywords:
Re-Planning, Re-Planning, Risk avoidance, Stochastic Dynamic Programming, Evasion, Roadmap, Risk analysis, Machine Forethought, UMAA

The US Navy has embarked on an effort to diversify its fleet to include fully unmanned surface platforms. As the endurance of these unmanned systems extends from days to months, pre-mission planning will be insufficient to handle the uncertainty in operations that occurs when transiting thousands of miles without operator input. Nowhere is this more pressing than in weather avoidance, where the chaotic nature of the dynamics limits practical forecasting to sixteen days of validity. Advanced planning and re-planning algorithms enable the platform to use machine forethought to avoid entering situations with dangerously restricted future actions. In Phase II, Metron developed a prototype of the Adverse Weather Avoidance, Re-planning, and Evasion (AWARE) planning tool that operates at the furthest extent of available meteorological and oceanographic (METOC) forecast data to enable long term planning. AWARE uses modern ideas in automated planning, synthesized forecast weather data, and incorporates platform capabilities to formulate a careful risk-aware decision process to map positions and times to the most favorable route. The resulting capability maximizes the opportunity for mission success while simultaneously minimizing risks to the platform. In Phase II.5, Metron will mature the prototype AWARE planning tool into a fully developed suite of software components compliant with the Navys Unmanned Maritime Autonomy Architecture (UMAA). Development will focus on implementing the UMAA services identified by Metron as dependencies of AWARE. To facilitate integration with the Navys fleet of USVs Metron will initially target a shoreside implementation of the software in coordination with the Unmanned Operations Center (UOC) located at NIWC-PAC in San Diego, California to perform trials and experimentation of prototype software components. As the AWARE suite is purely a software product that will follow the UMAA standards, the technical work involved in integrating with the UOC can be directly transferable to the Overlord-, Medium-, and Large-USV platforms once they become more widely available. Metron will work with the UOC and other third-party organizations to supply the UMAA services that are associated with platform specific hardware and executive autonomy functions. The integration with the UOC will happen in two phases: First, the UMAA services developed during Phase II will be integrated to quickly generate real world lessons learned to guide development during the remainder of this effort. Second, all the UMAA services developed during Phases II and II.5 will be integrated to demonstrate the full system capabilities.

Benefit:
The target of the technology proposed in this research is the M- and L- class USVs that currently being designed and built by the U.S. Navy. The AWARE planning tool fills an operational need to perform long term planning against dynamic and uncertain forecast data enabling unmanned systems to operate for thousands of miles without input from operators. The Phase II.5 product will follow the UMAA architecture specifications currently under development to ensure the AWARE planning tool can operate as a service on a DON vessel as part of an integrated autonomous system. The resulting Phase II.5 product will provide multiple avenues of transition to U.S. Navy systems as either an afloat instance on a fully autonomous system or shoreside as part of a Human in the Loop (HITL) system. Using either the afloat or the shoreside instance of AWARE enables route recommendations and risk assessments that are updated as new METOC data is received.

Keywords:
Route Generation, mission planning, Heatmap, Weather Avoidance, Risk avoidance, Autonomy, UMAA