Safely Landing Autonomous UAVs on Naval Vessels in Emergency Scenarios

Abstract

This project is for basic research that can improve the landing an autonomous lightweight unmanned air vehicle (UAV) on naval vessels. In particular, UMD is interested in dealing with things like the unpredictable motion of the landing area, with respect to the UAV, wind gusts, and maximizing the safety of people and equipment in the event of an unexpected motor-out emergency. To address these issues, UMD aims to combine a number of ideas from control and sampling based motion planning. Quickly generating a new motion plan, in this case for landing, is known as replanning. While it is always possible to recompute a new plan from scratch, such brute force replanning can be so time consuming that it is not practical in real-time emergency situations. Previous work on quick motion replanning has considered plans that respect the kinematics and dynamics of a vehicle, but has not considered the limits of the vehicle s closed loop controller during the regeneration of the motion plan leading to the possibility that the controller may be unable to follow the desired plan. On the other hand, methods do exist that are able to consider the vehicle’s closed loop control during the initial and possibly time consuming generation of the original plan. The proposed research is aimed at combining these ideas. UMD believes that explicitly accounting for the closed loop control within quick replanning algorithms will help to make the use of autonomous vehicles safer and more robust, especially in emergency situations.

Document Details

Document Type
DoD Grant Award
Publication Date
Dec 15, 2020
Source ID
N004212110001

Entities

People

  • Michael Otte

Organizations

  • United States Navy
  • University of Maryland

Tags

Readers

  • Educational Psychology
  • Robotics and Automation.
  • Systems Analysis and Design

Technology Areas

  • Autonomy
  • Autonomy - Autonomous System Control