Flight Testing of Intelligent Motion Video Guidance for Unmanned Air System Ground Target Surveillance Conference Paper uri icon


  • © 2018 by Noren, Valasek, Goecks, Rogers, Bowden. The need for an intelligent Small Unmanned Air System to assist in the surveillance and tracking of ground targets of interest has spurred the development of unconventional solutions. The challenge of operating an Unmanned Air System with a non-gimbaled or fixed camera increases operator workload since the vehicle must be steered to visually track targets. This paper details the implementation and initial fight testing of a machine learning algorithm for the autonomous tracking of ground targets. The Reinforcement Learning agent uses the Q-Learning algorithm and learns a control policy to keep a target within the camera image frame without user intervention. It is implemented onboard the vehicle rather than the ground control station, and functions as an outer-loop controller that commands bank angle increments to the autopilot. Prior to flight testing the Reinforcement Learning agent is trained offine in a simulation environment and learned different control policies to successfully track targets based upon target trajectories and crosswinds. The system architecture and challenges for the transition from a simulation to a hardware demonstration are presented. Based upon the flight test results presented in the paper, the approach is shown to be effective for tracking ground targets and is judged to be a candidate for an Unmanned Air System ground target tracking system.

author list (cited authors)

  • Noren, C., Valasek, J., Goecks, V. G., Rogers, C., & Bowden, E.

publication date

  • January 1, 2018 11:11 AM