SBIR/STTR Award attributes
The objective of this project is to develop a sense and avoid perception system for unmanned underwater vessels (UUVs) to support safe maneuvering and navigation in both the surface and the undersea domains. The market for UUVs has been growing rapidly, and with it, demand for obstacle avoidance systems. Growth is happening in the defense, commercial, and scientific sectors. It is worth acknowledging upfront that safely bringing an underwater vehicle to the surface is a complex problem. Human beings aboard submarines equipped with large, sophisticated sensors go through years of training to understand how to interpret sonar data in order to develop situational awareness on the surface. Automating this workflow requires a deep understanding of the sensors and the algorithms involved at each stage of the processing chain, how human operators make decisions, and an awareness of regulations and procedures, like COLREGs. For decades, Metron has been developing and transitioning advanced sonar and data fusion algorithms to the submarine fleet. Over the past decade, Metron has built large UUVs outfitted with our custom sensor systems and prototype collision avoidance algorithms. In recent years, we have made significant strides in applying automation at each stage of the sonar processing chain and have been an industry pioneer in obstacle avoidance algorithms for water-borne vehicles. Given the myriad of UUV designs that exist, the sought-after collision avoidance system must be able to work with different sensor configurations to be generally useful. This proposal outlines a plan to leverage some of Metron's successes to create an extensible multi-sensor data fusion guidance platform called RASCAL (Reconfigurable Autonomous Sensing and Collision Avoidance pLatform). During Phase II, a prototype multi-sensor hardware system will be built for small-class UUVs. RASCAL will encompass a set of algorithms with clearly defined interfaces to this prototype hardware and to the other UMAA components. The long-term vision for RASCAL is to become the surfacing and collision avoidance service within UMAA with specification-defined interfaces to sensing hardware using industry standards whenever possible. RASCAL’s algorithms will be designed from the ground up to work with different sensor configurations. Further, UI tools can be provided to assist UUV designers with insight into the collision avoidance performance implications of varying sensor placements.
