Distributed Control and Vision-Based Estimation for UAS Autonomy

Abstract

This research will address the problem of coordinating multiple unmanned aircraft systems (UAS) to autonomously achieve and maintain a desired formation using vision to estimate relative position between UASes and estimate their motion in the environment. There are two primary research objectives. The first objective is to make formation control more robust to loss or jamming of global measurements or information (e.g., GPS signal and connection to the command center) and require no communication between UASes. This will be achieved by developing distributed controllers that need only inter-vehicle relative position measurements and that work under dynamic sensing topology. Stability of the formation will be rigorously established using theories of control and dynamic systems. The second objective is to use novel image-based pose estimation technique to enable UASes to estimate the relative position of their neighbors using only the passive vision sensor measurements. With the ability to achieve the desired formation in radio silence, the communication channel is free for other purposes, and UASes can increase their stealth. When other sensors such as GPS, IMU, etc. are available, motion estimated from the vision data can be fused with these measurements to attenuate the noise and improve the localization accuracy and make the agents robust to intermittent loss and interruption in sensor measurements.

Document Details

Document Type
DoD Grant Award
Publication Date
Apr 28, 2017
Source ID
FA86511710001

Entities

People

  • Nicholas Gans

Organizations

  • Air Force Research Laboratory
  • United States Air Force
  • University of Texas at Dallas

Tags

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Geospatial Intelligence and Artificial Intelligence Analytics
  • Robotics and Automation.

Technology Areas

  • Autonomy
  • Autonomy - Autonomous System Control
  • Space
  • Space - Spacecraft Maneuvers