DRIB—Distributed Robotic Intelligent Behaviors
This ATC lead Air Force SBIR Phase II program developed a new software architecture for teams of heterogeneous robots that allows ad hoc networking, distributed processing, proxy processing, and cooperative task execution. Working closely with the University of Minnesota ATC develop the DRIB software architecture for implementation upon future small, resource limited robot, sensor, and micro-platforms.
Recent military operations in Afghanistan and Iraq highlighted the need for U.S. forces to be able to find and defeat "deep-targets". Deep-targets are those that are behind enemy lines, and can be concealed, strongly guarded, highly mobile, or even buried. Examples include underground hardened bunkers, cave/tunnel systems, and highly mobile SCUD launchers. They are very hard to defeat with conventional weapons without causing tremendous collateral damage. Current military weapons and delivery systems have difficulty in finding, identifying, collecting data on, and defeating deep-targets. Even precision guided bombs require accurate information on the target's location, composition, and protection. Thus, innovative approaches are needed to address these challenging targets.
The Air Force has identified the use of miniature or micro-robotics to perform deep-target reconnaissance and defeat missions as a promising way to effectively find and defeat "deep-targets". The envisioned scenario would involve dropping a number (perhaps hundreds) of small robots into a suspected target area. A variety of different types of robots with differing payloads and sensors would be employed. The robots would be small (few inches in diameter), and inexpensive to produce. If some are destroyed or lost when dropped (or during the mission) it would not be costly (in both dollars and in the mission's success). Enough robots would be dropped in the target area to allow the mission to continue regardless of depletion levels.
In order to support this envisioned scenario, ATC in conjunction with the University of Minnesota's Distributed Robotics Laboratory, developed the technology needed to support use of a heterogeneous cluster of inexpensive, miniature robotics to perform deep-target reconnaissance missions. The Distributed Robotic Intelligent Behavior (DRIB) system architecture and its associated intelligent behaviors provide:
- An agent based framework that provides the underlying support functionality that enables a large numbers of heterogeneous robots to coordinate their activity as they perform deep-target reconnaissance missions.
- Intelligent behaviors that enable the robots to carry out their missions
Our approach allows for the configuring and reconfiguring of the individual robots to execute assigned behaviors. DRIB is built on:
- ATC distributed robotics technology
- Research related to distributed robotics, coordination between autonomous entities, and miniature robotic platforms being conducted by Drs. Papanikolopoulos and Gini of the University of Minnesota.