Files in this item



application/pdfOrtizRubiano_Andres.pdf (4MB)
(no description provided)PDF


Title:On-line task scheduling and trajectory planning techniques for reconnaissance missions with multiple unmanned aerial vehicles supervised by a single human operator
Author(s):Ortiz Rubiano, Andres
Director of Research:Langbort, Cedric
Doctoral Committee Chair(s):Langbort, Cedric
Doctoral Committee Member(s):Voulgaris, Petros G.; Kirlik, Alex; Frazzoli, Emilio
Department / Program:Aerospace Engineering
Discipline:Aerospace Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):human machine interfaces
unmanned aircraft
trajectory planning
linear programming
task scheduling
Abstract:The problem of a single human operator monitoring multiple UAVs in reconnaissance missions is addressed in this work. In such missions, the operator inspects and classifies targets as they appear on video feeds from the various UAVs. In parallel, the aircraft autonomously execute a flight plan and transmit real-time video of an unknown terrain. The main contribution of this work is the development of a system that autonomously schedules the display of video feeds such that the human operator is able to inspect each target in real time (i.e., no video data is recorded and queued for later inspection). The construction of this non-overlapping schedule is made possible by commanding changes to the flight plan of the UAVs. These changes are constructed such that the impact on the mission time is minimized. The development of this system is addressed in the context of both fixed and arbitrary target inspection times. Under the assumption that the inspection time is constant, a Linear Program (LP) formulation is used to optimally solve the display scheduling problem in the time domain. The LP solution is implemented in the space domain via velocity and trajectory modifications to the flight plan of the UAVs. An online algorithm is proposed to resolve scheduling conflicts between multiple video feeds as targets are discovered by the UAVs. Properties of this algorithm are studied to develop conflict resolution strategies that ensure correctness regardless of the target placement. The effect of such strategies on the mission time is evaluated via numerical simulations. In the context of arbitrary inspection time, the human operator indicates the end of target inspection in real time. A set of maneuvers is devised that enable the operator to inspect each target uninterruptedly and indefinitely. In addition, a cuing mechanism is proposed to increase the situational awareness of the operator and potentially reduce the inspection times. The benefits of operator cuing on mission time are evaluated through a numerical study using various levels of simulated human response.
Issue Date:2012-09-18
Rights Information:Copyright Andres Eduardo Ortiz Rubiano. All rights reserved
Date Available in IDEALS:2012-09-18
Date Deposited:2012-08

This item appears in the following Collection(s)

Item Statistics