Return to search

Path Planning for a UAV in an Agricultural Environment to Tour and Cover Multiple Neighborhoods

This work focuses on path planning for an autonomous UAV to tour and cover multiple regions in the shortest time. The three challenges to be solved are - finding the right optimal order to tour the neighborhoods, determining entry and exit points to neighborhoods, and covering each neighborhood. Two approaches have been explored and compared to achieve this goal - a TSP - Greedy and TSP - Dijkstra's. Both of them use a TSP solution to determine the optimal order of touring. They also use the same back and forth motion to cover each region. However, while the first approach uses a brute force to determine the the next closest node of entry or exit, the second approach utilizes the Dijkstra's algorithm to compute all possible paths to every node in the graph, and therefore choose the shortest pairs of entry and exit for each region, that would generate the shorter path, overall. The main contribution of this work is to implement an existing algorithm to combine the touring and covering problem, and propose a new algorithm to perform the same. Empirical results comparing performances of both approaches are included. Hardware experiments are performed on a spraying hexacopter, using the TSP - Greedy approach. Unique system characteristics are studied to make conclusions about stability of the platform. Future directions are identified to improve both software and hardware performance. / Master of Science / In a world with a rapidly growing population and resources depleting faster, increasing efficiency and productivity has become paramount. Until now, automation has helped cope with the world’s increasing demand for food. However, studies have shown that automation in itself will be insufficient in improving crop output in the coming years. Fortunately, another technology that is taking big leaps in terms of technological advances - Information Technology, when combined with automation, presents itself as a viable option. This takes agriculture towards a site-specific approach for all crop monitoring, growth and protection activities and is know as Precision Agriculture. Spraying fluids on crops using a UAV is on of the prominent problems being researched in this field. This work presents two approaches - TSP - Greedy and TSP - Dijkstra’s to tour or visit and spray multiple regions that have been previously identified in the shortest time. While the TSP - Greedy algorithm is an implementation of an existing approach, the TSP - Dijkstra’s algorithm is a new approach proposed in this work. The solution to TSP or Traveling Salesman Problem generates the optimal order to visit the regions. The ’Greedy’ or ’Dijsktra’s’ approaches define entry and exit points for each region, that gives the shortest path overall. Images of areas with weed afflicted regions marked on them are used as the input for this algorithm. The TSP-Greedy approach is used in performing hardware experiments. Data collected from these experiments are used to analyze performance of an Unmanned Aerial Vehicle (UAV) platform. Water has been used as the spraying fluid for testing the sprayer assembly. GPS or Global Positioning System is used for navigation of the UAV. Future directions are identified to improve both software and hardware performance.

Identiferoai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/79731
Date20 October 2017
CreatorsSinha, Koel
ContributorsElectrical and Computer Engineering, Kochersberger, Kevin B., Tokekar, Pratap, Roan, Michael J.
PublisherVirginia Tech
Source SetsVirginia Tech Theses and Dissertation
Detected LanguageEnglish
TypeThesis
FormatETD, application/pdf
RightsIn Copyright, http://rightsstatements.org/vocab/InC/1.0/

Page generated in 0.0029 seconds