Autonomus Mission Execution for Teams of Reconnaissance UAVs

Abstract

One of the key missions for reconnaissance UASs at sea is to locate vessels, people, andother objects as quickly as possible with the limited flight time of the vehicles. In the currentparadigm the vehicle streams video to an operator which tasks the vehicle with fixed missionsequences or commands. This way of operating the vehicles is not effective for several reasons.Maintaining constant communication with the base station can be difficult especially at longerranges, and increases vulnerability for the base station and vehicle to be found. Additionally,as the number of vehicles increases the bandwidth used increases linearly. Since each vehiclerequires at least one operator to supervise the video stream, and the number of vehicles thatcan be used simultaneously is limited by the number of base stations and operators.Instead, the goal in this project is to give the human operator the ability to send high levelmissions and constraints, replace the constant human supervision with automateddetections, and automate decision making on the vehicle with an autonomy system. In theproposed project we will attempt to developed a mission autonomy system that fuses theonboard detections, shares coordination information with other vehicles, maximizes relevantinformation gathered based on the information available, and learns how to make robustdecision based on the information available. We will evaluate the system in several simulation experiments of increasing difficulty and expect to be able to demonstrate the effectiveness of our method to reduce the communication requirements and increase search efficiency.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 06, 2021
Source ID
N000142112110

Entities

People

  • Sebastian Scherer

Organizations

  • Carnegie Mellon University
  • Office of Naval Research
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Applied Combinatorial Optimization and Logic Circuit Design.
  • Neural Network Machine Learning.
  • Unmanned Aerial System (UAS) Autonomous Capabilities and Mission Reconnaissance.