Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning
| Main Author: | |
|---|---|
| Language: | English |
| Published: |
The Ohio State University / OhioLINK
2020
|
| Subjects: | |
| Online Access: | http://rave.ohiolink.edu/etdc/view?acc_num=osu1587074034916099 |
| id |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu1587074034916099 |
|---|---|
| record_format |
oai_dc |
| spelling |
ndltd-OhioLink-oai-etd.ohiolink.edu-osu15870740349160992021-08-03T07:14:29Z Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning Soderlund, Alexander A. Aerospace Engineering Wildfire estimation information fusion The objective of this dissertation is to develop a bi-directional feedback framework that autonomously estimates the state of an evolving wildland fire in real-time. The fire is cast as a random dynamical system whose time-varying physical states evolve according to stochastic environmental conditions. The state estimation described in this work is adaptive in that we employ a sensing set that is reconfigurable based on the simulated physical process. The estimated fire's trajectory provides feedback to the sensing set by guiding their deployment to advantageous locations in order to enhance the information content of their measurements. The sensing set in turn provides feedback via higher-quality observations thereby improving the precision of the wildfire estimate and completing the bi-directional architecture.Conventionally, the state estimation process of updating the forecasted dynamics with available measurements is done via Bayesian conditioning on the premise that an adequate prior probability distribution is available. This work alternatively explores information fusion within the framework of Dempster-Shafer theory which does not rely on the certainty of a Bayesian prior, but rather combines agents' beliefs based on the pooling of available evidence. Fusion is performed through Dempster's rule of combination which incorporates each agent's degree of ignorance regarding the evidence into a combined posterior belief. This act of fusion additionally provides an inter-agent conflict value that may indicate faulty beliefs on the part of one (or multiple) agents.In the context of the wildfire scenario, three types of agents are considered whose beliefs are respectively formed from the following distinct bodies of evidence: (i) physics-driven fire spread forecasts, (ii) location-specific temperature readings, and (iii) visible imagery of the environment. These belief models were derived from a synthesis of fire dynamics and empirical data gathered from a prescribed large-scale fire test. The efficacy of the estimation procedure is demonstrated over two numerically simulated wildfire scenarios under conditions of varying environmental uncertainty. 2020-10-01 English text The Ohio State University / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=osu1587074034916099 http://rave.ohiolink.edu/etdc/view?acc_num=osu1587074034916099 unrestricted This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws. |
| collection |
NDLTD |
| language |
English |
| sources |
NDLTD |
| topic |
Aerospace Engineering Wildfire estimation information fusion |
| spellingShingle |
Aerospace Engineering Wildfire estimation information fusion Soderlund, Alexander A. Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| author |
Soderlund, Alexander A. |
| author_facet |
Soderlund, Alexander A. |
| author_sort |
Soderlund, Alexander A. |
| title |
Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| title_short |
Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| title_full |
Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| title_fullStr |
Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| title_full_unstemmed |
Characterization of Wildland Fires through Evidence-basedSensor Fusion and Planning |
| title_sort |
characterization of wildland fires through evidence-basedsensor fusion and planning |
| publisher |
The Ohio State University / OhioLINK |
| publishDate |
2020 |
| url |
http://rave.ohiolink.edu/etdc/view?acc_num=osu1587074034916099 |
| work_keys_str_mv |
AT soderlundalexandera characterizationofwildlandfiresthroughevidencebasedsensorfusionandplanning |
| _version_ |
1719457178224427008 |