Experimental Results of Using Swarming Techniques for Real-time Tasking of UAVs

Carolyn Boettcher

This talk will describe the results of applying swarming algorithms to provide a real-time solution to the problem of assigning UAVs to geographically dispersed targets and scenes. The problem is defined by the number of vehicles and locations involved. Although there are other methods that provide optimal or near-optimal solutions to this problem, they become intractable as the size of the problem grows. Our results show that swarming gives superior results for problems with hundreds of assets and tasks, which are not practical to solve with other search techniques. The talk discusses results obtained from thousands of Monte Carlo simulations and compares swarming to results from other algorithms. It also describes some of the real life practical constraints that need to be considered in the solution.

Audio (MP3 File, Podcast Ready) Presentation Files (Zip Archive)

Back to Swarming by Nature and by Design