טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentMoran Yehuda
SubjectInfo-Gap Approach for Micro UAV Helicopter Motion Planning
under Severe Uncertainty
DepartmentDepartment of Aerospace Engineering
Supervisors Professor Tal Shima
Professor Yakov Ben-Haim
Full Thesis textFull thesis text - English Version


Abstract

A robust-satisficing approach, based on the Info-Gap Decision Theory, is suggested for path assessment of unmanned aerial vehicles. In this study, an unmanned helicopter, using a gimballed sensor with a limited field of view, tracks the severely uncertain motion of two cars within a finite time horizon. It is assumed that the tracking mission may be performed in an open terrain in which the cars choose any pattern that satisfies their motion constraints, or, in a constrained environment in which the cars choose a road out of a limited set of possibilities. Both models take into account a priori information regarding the cars’ movement, but such a priori information is assumed to be severely uncertain. A methodology is suggested for real-time selection of a robust-satisficing path for the helicopter, taking into account applicable motion and visibility constraints, and the targets' motion uncertainties. The robust-satisficing path within the constrained environment model is obtained by analytic solution, however, because of the burden involved in the complex computations that are required to solve the open terrain problem, a Genetic Algorithm is proposed. A performance analysis was conducted to study the effect of the problem's parameters over the robustness of the helicopter's path. The results were compared to the helicopter's optimal path. Subsequently, a Monte Carlo study was conducted to show the relation between the helicopter's flight patterns and its robustness to uncertainty. From the results it is evident that: a) In various scenarios the robust-satisficing path meets the required performance criteria while tolerating higher levels of uncertainty in the a priori information in relation to the optimal path; b) Robustness increases as the helicopter's field of view and the mission suspension time increase; and c) In order to achieve higher robustness, the helicopter is required to increase deviation from the nominal paths computed based on the a priori information concerning anticipated paths of the targets. The simulation results were tested in the cooperative autonomous systems laboratory. The flight test showed that due to uncertainty in the helicopter's dynamic model it failed to properly track the desired path. Although initially the helicopter's motion uncertainty wasn't considered in this research the results of the flight test showed it is not negligible. In order to immunize the decision maker to both uncertainties in the targets' and the helicopter's motions an Info-Gap model considering the uncertainty in the motion of the helicopter should be added.