Publicação
Radar Emitter Classification based on Deep Ensemble
| datacite.subject.fos | Ciências Naturais::Ciências da Computação e da Informação | pt_PT |
| dc.contributor.advisor | Henriques, Roberto André Pereira | |
| dc.contributor.author | Martins, Tiago Pedro Giesta | |
| dc.date.accessioned | 2023-11-24T16:52:33Z | |
| dc.date.available | 2023-11-24T16:52:33Z | |
| dc.date.issued | 2023-10-26 | |
| dc.description | Dissertation presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence | pt_PT |
| dc.description.abstract | Electronic Support Measures (ESM) systems are designed to classify radar signals, providing information about the presence of threats. This function aids in battlefield situational awareness and the commander's decision on which countermeasures to employ. This dissertation aims to develop a deep ensemble model, recognizing the importance of a fast and precise classification based on a deep forest as an alternative to the parameter matching method. Four deep ensemble models and six of its base learners were built and evaluated to classify 52 emitters, using seven train/test datasets and two test datasets with noise, totalling 420 measurements of accuracy and classification speed. After analyzing these results, two deep ensemble models and their base learners were optimized, each for a different dataset, achieving 100% accuracy in a feature-engineered dataset and up to 98.358% in the original dataset. Regarding classification speed, the fastest models can classify 1000 records in 64ms, which may be acceptable in the real world. The experimental results of this approach reveal several advantages, making it a feasible alternative, including reduced dependency on ESM experts, ease of maintenance, quick to update, and high accuracy. | pt_PT |
| dc.identifier.tid | 203391470 | pt_PT |
| dc.identifier.uri | http://hdl.handle.net/10362/160450 | |
| dc.language.iso | eng | pt_PT |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | pt_PT |
| dc.subject | Ensemble methods | pt_PT |
| dc.subject | Deep architecture | pt_PT |
| dc.subject | Electronic Support Measures | pt_PT |
| dc.subject | Classification | pt_PT |
| dc.subject | SDG 8 - Decent work and economic growth | pt_PT |
| dc.subject | SDG 9 - Industry, innovation and infrastructure | pt_PT |
| dc.subject | SDG 17 - Partnerships for the goals | pt_PT |
| dc.title | Radar Emitter Classification based on Deep Ensemble | pt_PT |
| dc.type | master thesis | |
| dspace.entity.type | Publication | |
| rcaap.rights | openAccess | pt_PT |
| rcaap.type | masterThesis | pt_PT |
| thesis.degree.name | Mestrado em Gestão de Informação, especialização em Gestão do Conhecimento e Inteligência de Negócio (Business Intelligence) | pt_PT |