UAV Position Optimization for Wireless Communications
Abstract
This thesis explores autonomously positioning unmanned aerial vehicles (UAV) as wireless nodes in optimal locations to form robust, reliable communication links between static or slow-moving nodes, on land or at sea, in a wireless network. The presented approach explicitly accounts for variability associated with signal-to-noise ratio (SNR) estimates used for UAV navigation. A two-phased approach is presented to find a local SNR extremum as an optimal loitering point. This thesis focuses on phase one consisting of Kriging and semivariogram analysis as well as information theoretic local path planning. Kullback-Leibler divergence is used for path evaluation and selection. Phase two consists of an extremum control method developed in prior work for UAV navigation to the optimal loitering point. Emphasis is placed on accuracy and reducing model uncertainty. Simulated and experimental data is presented and used for Kriging of the SNR field produced by two ground nodes. Datasets produced with varying distances, altitudes, and flight patterns provide insight into the behavior of SNR degradation and flight trajectories that are most efficient at reducing estimate uncertainty. Analysis provides a greater understanding of the current capabilities, benefits, and limitations of employing UAVs as autonomous, mobile communication nodes. This includes the potential for implementing nonlinear optimal estimation and path planning processes onboard small UAVs in real time.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2018
- Accession Number
- AD1059967
Entities
People
- Benjamin P. Keegan
Organizations
- Naval Postgraduate School