Linear Regression Models Applied to Imperfect Information Spacecraft Pursuit-Evasion Differential Games

Abstract

Within satellite rendezvous and proximity operations lies pursuit-evasion differential games between two spacecraft. The extent of possible outcomes can be mathematically bounded by differential games where each player employs optimal strategies. A linear regression model is developed from a large data set of optimal control solutions. The model is shown to map pursuer relative starting positions to final capture positions and estimate capture time. The model is 3.8 times faster than the indirect heuristic method for arbitrary pursuer starting positions on an initial relative orbit about the evader. The linear regression model is shown to be well suited for on-board implementation for autonomous mission planning.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 26, 2020
Accession Number
AD1101535

Entities

People

  • Dax A Linville

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Space

DTIC Thesaurus Topics

  • Air Force
  • Closed Loop Systems
  • Computer Programming
  • Computer Programs
  • Computers
  • Control Systems
  • Coordinate Systems
  • Department Of Defense
  • Differential Equations
  • Equations Of Motion
  • Heuristic Methods
  • Kalman Filters
  • Mathematical Filters
  • Orbital Elements
  • Particle Swarm Optimization
  • Relative Motion
  • Spacecraft

Fields of Study

  • Mathematics

Readers

  • Game Theory.
  • Robotics and Automation.
  • Statistical inference.

Technology Areas

  • Space
  • Space - Spacecraft Maneuvers