ML-Recon simulation model: a Monte Carlo planning aid for Magic Lantern.

Abstract

The U.S. Navy currently has no means to conduct sea mine reconnaissance with assets that are organic to Aircraft Carrier Battle Groups or Amphibious Ready Groups. Magic Lanten is an Airborne Laser Mine Detection System (ALMDS) under development, that is designed to search for floating and shallow moored mines using a helicopter- mounted laser-optic sensor. It is the only ALMDS operationally tested by the Navy to date. This thesis develops a Monte Carlo simulation model called ML-Recon, which is intended for use as a tool to plan mine reconnaissance searches using the Magic Lantern system. By entering fundamental initial planning information, the user can determine the number of uniformly-spaced tracks to fly with a Magic Lantern-equipped helicopter to achieve a certain level of assurance that the area contains no floating or shallow moored mines. By employing Monte Carlo methods, ML-Recon models the three primary stochastic processes that take place during a typical search: the location of the mines, the cross-track error of the helicopter, and the detection/non-detection process of the sensor. By running ML-Recon, the user is given performance statistics for many replications of the search plan that he chooses. This approach is unique in that it provides the user with information indicating how much worse than the mean performance his plan may perform. ML- Recon also gives the user an Opportunity to view an animation of his lan, which he can use to look for tendencies in the lan to contain holes, or holidays. (AN)