Prev-Talk | Next-Talk | All-Talks | Talks-Sorted
Talk-09: ''Information Theory Based Multi-Vehicle Collaboration for Multi-Static Sonar Using MOOS-IvP"
Mike Hamilton, Stephanie Kemna, David Hughes, NATO Undersea Research Centre (NURC), Italy
Within the Cooperative Anti-submarine Warfare (CASW) project at the NATO Undersea Research Centre (NURC), a set of behaviors has been developed to optimize multi-static sonar performance. This is ac- complished using NURC’s AUV towing a sonar receiver array. The vehicle maneuvers to position the array in the most favorable position for local- izing a detected target in a bistatic system. We refer to this as our optimization behavior, as it is a behavior which optimizes the localization of the target.
This optimization behavior has been extended to multiple vehicles. Using a two vehicle system, the first vehicle performs the standard optimization behavior. The second vehicle, using information from the first, moves not to optimize his own performance, but to optimize the performance of the entire system. An information theory based approach is used as the optimization function. A one-step look ahead implementation has been completed with MOOS-IvP and demonstrated in the recent GLINT10 sea trial. Interesting aspects in the differences between the joint and individual localization approach will be described. Also, extensions to a more general receding horizon approach, and corresponding MOOS-IvP implementation challenges, will be discussed.
Related Material:
Categories:
- Autonomy
- MOOS-IvP
- Multi-Vehicle Autonomy
- ASW
- UUVs
- Govt Labs