Optimal Control of an Uninhabited Loyal Wingman
Abstract
This work demonstrates the use of optimal control and stochastic estimation techniques as an autonomous near real-time dynamic route planner for the DoD concept of the loyal wingman. First, the optimal control problem is formulated for a static threat environment and a hybrid numerical method is demonstrated. The optimal control problem is transcribed to a nonlinear program using direct orthogonal collocation, and a heuristic particle swarm optimization algorithm is used to supply an initial guess to the gradient-based nonlinear programming solver. Next, a dynamic and measurement update model and Kalmanfilter estimating tool is used to solve the loyal wingman optimal control problem in the presence of moving, stochastic threats. Finally, an algorithm is written to determine if and when the loyal wingman should dynamically re-plan the trajectory based on a modified distance to re-plan formulation which uses speed and stochastics of the moving threat as well as relative distance and angle of approach of the loyal wingman to the threat. Results demonstrate a methodology for rapidly computing an optimal solution to the loyal wingman optimal control problem.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2016
- Accession Number
- AD1018255
Entities
People
- Clay J. Humphreys
Organizations
- Air Force Institute of Technology