Autonomous vehicles have many applications in tasks that are too dangerous or mundane for human operators.  The cooperative control of teams of such vehicles is one approach to enhancing the capabilities of these vehicles.  One example of this is the coordination of teams of wide-area search munitions.  These munitions will be able to autonomously (with minimal operator oversight) search an area, identify targets, and conduct an attack.

The following movies illustrate a concept for cooperative attack using sensor information from multiple munitions to perform target-location estimation while carrying out an attack.  The top-left and bottom-right views are static views of the target: a surface-to-air missile launcher.  The top-right and bottom-left views are views from two munitions.  The top-right munition carries out the attack, and the bottom-left munition collects sensor information to aid in the target-location estimation.  The munitions follow specially designed Simulaneous Localization and Planning (SLAP) trajectories that help reduce the uncertainty in the target location.

(Note: These movies were created using the Indeo codec.  To play in Windows Media Player they need to be saved to your hard drive.  Also, other people have been able to play these movies without difficulty, but I actually have to turn down hardware acceleration in the display settings on my machine for the colors to display correctly.)

It has been shown that using the SLAP trajectories can improve target-location estimation, reducing the location uncertainty to one quarter of the area achieved when using naively designed trajectories.  Research is continuing to allow the implementation of SLAP trajectories in a real-time architecture.

This work has been in collaboration with Chad Prazenica (University of Florida, REEF) and David Jeffcoat (Air Force Research Lab, Eglin AFB).