Online Unmanned Ground Vehicle Mission Planning using Active Aerial Vehicle Exploration

Simple item page
dc.contributor.authorWagner, Anthony Julianen
dc.contributor.committeechairKochersberger, Kevin B.en
dc.contributor.committeememberTokekar, Pratapen
dc.contributor.committeememberWicks, Alfred L.en
dc.contributor.departmentMechanical Engineeringen
dc.date.accessioned2019-06-29T08:01:19Zen
dc.date.available2019-06-29T08:01:19Zen
dc.date.issued2019-06-28en
dc.description.abstractThis work presents a framework for the exploration and path planning for a collaborative UAV and UGV system. The system is composed of a UAV with a stereo system for obstacle detection and a UGV with no sensors for obstacle detection. Two exploration algorithms were developed to guide the exploration of the UAV. Both identify frontiers for exploration with the Dijkstra Frontier method using Dijkstra's Algorithm to identify a frontier with unknown space, and the other uses a bi-directional RRT to identify multiple frontiers for selection. The final algorithm developed was for to give the UGV partial plans when an entire plan is not yet found. This improves the overall mission tempo. The algorithm is designed to keep the UGV a safe distance from the unknown frontier to prevent backtracking. All the algorithms were tested in Gazebo using the ROS framework. The Dijkstra Frontier method was also tested on the hardware system. The results show the RRT Explore algorithm to work well for exploring the environment, performing equally or better than the Dijkstra Frontier method. The UGV partial plan method showed a decreased traveled distance for the UGV but increases in UGV mission time with more conservative distances from danger. Overall, the framework showed a good exploration of the environment and performs the intended missions.en
dc.description.abstractgeneralThis work presents a framework for the exploration and path planning for a collaborative aerial and ground vehicle robotic system. The system is composed of an aircraft with a camera system for obstacle detection and a ground vehicle with no sensors for obstacle detection. Two exploration algorithms were developed to guide the exploration of the aircraft. Both identify frontiers for exploration with the Dijkstra Frontier method using path planning algorithms to identify a frontier with unknown space (Dijkstra Frontier), and the other uses a sampling based path planning method (RRT Explore) to identify multiple frontiers for selection. The final algorithm developed was for to give the ground vehicle intermediate plans when an entire plan is not yet found. The algorithm is designed to keep the ground vehicle a safe distance from the unknown frontier to prevent backtracking. All the algorithms were tested in a simulation framework using Robot Operating System and one exploration method was tested on the hardware system. The results show the RRT Explore algorithm to work well for exploring the environment, performing equally or better than the Dijkstra Frontier method. The ground vehicle intermediate plan method showed a decreased traveled distance for the ground vehicle but increases in ground vehicle mission time with more conservative distances from danger. Overall, the framework showed a good exploration of the environment and performs the intended missions.en
dc.description.degreeMaster of Scienceen
dc.format.mediumETDen
dc.identifier.othervt_gsexam:21485en
dc.identifier.urihttp://hdl.handle.net/10919/90785en
dc.publisherVirginia Techen
dc.rightsIn Copyrighten
dc.rights.urihttp://rightsstatements.org/vocab/InC/1.0/en
dc.subjectDrone aircraften
dc.subjectUGVen
dc.subjectExplorationen
dc.subjectRRTen
dc.titleOnline Unmanned Ground Vehicle Mission Planning using Active Aerial Vehicle Explorationen
dc.typeThesisen
thesis.degree.disciplineMechanical Engineeringen
thesis.degree.grantorVirginia Polytechnic Institute and State Universityen
thesis.degree.levelmastersen
thesis.degree.nameMaster of Scienceen
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Wagner_AJ_T_2019.pdf
Size:
25.31 MB
Format:
Adobe Portable Document Format
Download
Collections
Masters Theses