Robust online adaptive sensor-driven survey planning for single and multiple autonomous underwater vehicles
| dc.contributor.advisor | Li, Howard | |
| dc.contributor.advisor | Seto, Mae | |
| dc.contributor.advisor | Chang, Liuchen | |
| dc.contributor.author | Paull, Liam | |
| dc.date.accessioned | 2023-03-01T16:16:38Z | |
| dc.date.available | 2023-03-01T16:16:38Z | |
| dc.date.issued | 2013 | |
| dc.date.updated | 2018-10-29T00:00:00Z | |
| dc.description.abstract | The objective of the thesis is to propose new survey planning methods for autonomous underwater vehicles (AUVs). Historical approaches to the problem usually involve preplanning paths before the start of the survey in a structured, either zig-zag or lawn mower, pattern. This has three notable shortcomings: 1) The AUV may not be able to follow the prescribed path exactly, for example if there are unexpected disturbances, 2) Object detection in imagery obtained from the survey is dependent on many factors, some of which may not be known a priori, resulting in overly conservative plans, and 3) In the absence of pre-placed beacons, once the vehicle submerges, its location estimate will drift, again resulting in deviation from the pre-specified path. These issues are overcome here using new online and adaptive approaches to survey planning. First, an algorithm is developed that maintains a model of the area coverage over the workspace in real-time and plans paths within that workspace. This sensordriven algorithm is also able to account for uncertainty in the factors that affect object detection while underway. Second, the algorithm is augmented to explicitly account for the uncertainty in the A UV location estimate. The link between area coverage and state estimation is made. It is motivated that for coverage, estimating the full trajectory ( or smoothing) as opposed to just the present pose ( or filtering) is beneficial. In addition, the planning strategies previously developed are augmented to operate within the new probabilistic framework. The main benefit of the proposed approach is that it is robust to localization sensor noise and can guarantee coverage in a real field coverage sense not just a planning sense. Finally, the algorithms are extended to coverage with multiple AUVs. If a team of AUVs possess precisely synchronized clocks, then when they communicate underwater they can also calculate their relative ranges from the time-of-flight of the acoustic signals. A cooperative trajectory estimation algorithm is developed that is particularly well-suited to operate within the challenging underwater communications channel. Data to be transmitted in each broadcast communication scales linearly with the number of vehicles in the AUV team and does not backlog in the case of communication failures, which are common underwater. The algorithms are tested in simulation and in the field on the Iver2 AUV. These tests show the effectiveness of the proposed methods. | |
| dc.description.copyright | © Liam Paull, 2014 | |
| dc.format | text/xml | |
| dc.format.extent | xxi, 241 pages | |
| dc.format.medium | electronic | |
| dc.identifier.uri | https://unbscholar.lib.unb.ca/handle/1882/13246 | |
| dc.language.iso | en_CA | |
| dc.publisher | University of New Brunswick | |
| dc.rights | http://purl.org/coar/access_right/c_abf2 | |
| dc.subject.discipline | Electrical and Computer Engineering | |
| dc.title | Robust online adaptive sensor-driven survey planning for single and multiple autonomous underwater vehicles | |
| dc.type | doctoral thesis | |
| thesis.degree.discipline | Electrical and Computer Engineering | |
| thesis.degree.fullname | Doctor of Philosophy | |
| thesis.degree.grantor | University of New Brunswick | |
| thesis.degree.level | doctoral | |
| thesis.degree.name | Ph.D. |
Files
Original bundle
1 - 1 of 1