Online Optimization to Increase Small Unmanned Aerial Vehicle Surveillance Capacity in Joint Forcible Entry Operations
Abstract
This research develops and tests a treasure-collecting unmanned aerial vehicle (UAV) model that locates and observes enemy elements during reconnaissance. Operating in a simulated Joint Forcible Entry (JFE) battlefield, the model organizes its environment in discrete time and space. Using time-series forecasting, online optimization (OO), and multi-objective optimization (MOO), the UAV predicts enemy interception points while considering uncertainty. An experiment compared two use cases across 64 factor levels. The max-collect use case enabled the UAV to detect six times more enemies. Results showed that in enemy-scarce scenarios, the UAV was twice as likely to arrive late at the Named Area of Interest (NAI) compared to higher enemy presence scenarios. Longer transits posed computational challenges for prediction and optimization. Model tuning is necessary to balance performance, time cost, and computational energy. In conclusion, the experiment validated the models ability to trade UAV flight time for enhanced battlefield intelligence. These results are significant for preserving JFE capabilities in anti-access/area-denied (A2AD) environments. Future applications may involve multiple UAVs integrated with a common operational picture (COP) to supplement or replace existing large surveillance assets.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2023
- Accession Number
- AD1213510
Entities
People
- Herbert W. Jockheck
Organizations
- Naval Postgraduate School