Date of Award

2-2020

Document Type

Dissertation

Publisher

Santa Clara : Santa Clara University, 2020.

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

First Advisor

Christopher Kitts

Abstract

Scalar fields are spatial regions where each point has an associated physical value. These fields often contain features of interest, such as local extrema and contours with a value of significance. Traditional navigation techniques require robots to exhaustively search these regions to find the areas of significance, while adaptive navigation allows them to move directly to the points of interest based on measurements of the field taken during the navigation process. This work expands existing adaptive navigation techniques by adding a finite state machine layer to the control architecture, and using it as a discrete mode controller; the state machine allows for the sequencing of individual adaptive navigation control primitives for the purpose of enhancing performance and achieving new mission-level capabilities. For example, it has enabled improvements to existing ridge, trench, and saddle point navigators and the creation of a novel technique for navigating along scalar fronts. In both cases, experimental results demonstrated excellent tracking of the features of interest. Furthermore, mission-level capabilities were developed for low-exposure waypoint navigation and mapping contours round an extremum. These missions were evaluated through the use of 10,000 simulations with success rates of 96:95% for low exposure waypoint navigation and 87:36% for contour mapping.

Share

COinS