Investigações sobre modelagem, mitigação e predição de cintilação ionosférica na região brasileira
Carregando...
Arquivos
Data
2018-03-01
Autores
Vani, Bruno César [UNESP]
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Cintilações Ionosféricas são rápidas variações na amplitude e/ou fase de um sinal de rádio ao se propagar por irregularidades na densidade de elétrons na ionosfera. Este fenômeno degrada a performance do posicionamento pelo GNSS, uma vez que pode acarretar, dentre outros aspectos, a degradação na acurácia de observáveis e em perdas de sincronismo no receptor. No Brasil, verifica-se a maior suscetibilidade de ocorrência de cintilação algumas horas após o pôr-do-sol, nas épocas do ano compreendidas entre os equinócios de primavera e outono nos anos de alta atividade solar. Redes GNSS de monitoramento de cintilações estão implantadas no território brasileiro, incluindo a rede CIGALA/CALIBRA – a qual é mantida pela FCT/UNESP com o apoio de parceiros nos últimos sete anos. Os dados de monitoramento permitem a realização de diversas pesquisas sobre características e efeitos da cintilação, incluindo as investigações conduzidas neste projeto. Foram investigados aspectos sobre a modelagem da cintilação ionosférica no Brasil, com ênfase na mitigação dos seus efeitos no PPP e na predição de ocorrência de cintilação. No contexto da mitigação, abordagens existentes foram avaliadas e uma nova proposta foi desenvolvida. A abordagem proposta para mitigação consiste em um novo modelo funcional, novo modelo estocástico e uma estratégia para minimizar os efeitos de perdas de sincronismo. A abordagem proposta foi testada com o suporte do software científico RT-PPP e os resultados obtidos foram promissores, incluindo casos de recuperação da acurácia esperada do PPP, mesmo sob influência de cintilação forte. No contexto da predição, uma integração de bases de dados de monitoramento oriundos de três redes (CIGALA/CALIBRA, ICEA e LISN) permitiu o desenvolvimento de um modelo preditivo (guiado por dados) baseado em redes neurais artificiais. A rede neural é treinada para obter uma estimativa das localidades e horários onde a cintilação é esperada em uma determinada noite, com base em dados de monitoramento obtidos em noites anteriores. Dados de monitoramento de diferentes regiões coletados no começo da mesma noite (logo após o pôr-do-sol) também são utilizados com o objetivo de inferir padrões acerca do surgimento das irregularidades que causam as cintilações e sua relação com o nível de cintilação observado no restante da noite. O modelo permite obter mapas de cintilação preditos com antecedência de uma a quatro horas, os quais são acompanhados de estimativas de qualidade das predições. Em resumo, este projeto apresenta contribuições com potencial para trazer benefícios ao cenário científico-tecnológico nacional. Além disso, os dados de monitoramento da base de dados integrada foram disponibilizados pela internet à comunidade através do software científico ISMR Query Tool, proporcionando suporte à realização de pesquisas adicionais em diversas instituições do Brasil e do mundo.
Ionospheric scintillations are rapid variation in amplitude and/or phase of a radio signal as it propagates through irregularities on electron density in the ionosphere. Such phenomenon degrades the performance of GNSS positioning, because it may cause accuracy degradation on observables and losses of lock, among other aspects. In Brazil, there is more susceptibility to occurrence of scintillations after sunset time between the spring and autumn equinoxes of years with high solar activity. Monitoring networks based on GNSS receivers are deployed over the Brazilian territory such as the CIGALA/CALIBRA network, managed by FCT/UNESP (with support from partners) in the last seven years. Monitoring data allows to develop several research regarding the scintillation effects, as in this thesis. Aspects regarding the modeling of ionospheric scintillation effects in Brazil were investigated, with emphasis on the mitigation of these effects on PPP and predictions of scintillation occurence. In the field of mitigation, existing approaches were investigated and a new one was proposed. The new approach for mitigation relies in both new functional and stochastic models for PPP, as well as a strategy to model effects of losses of lock. The proposed approach was tested with the scientific software RT-PPP and the achieved results were promising, including cases in which the expected accuracy for the PPP was recovered. In the field of the predictions, a database integration was conducted with data from three different networks (CIGALA/CALIBRA, ICEA e LISN). The integration allowed the development of a data-driven predictive model based on artificial neural network. The neural network is trained with data from previous nights. Data from the same night (around the sunset time) is also used to detect patterns regarding the emerging of the irregularities driving scintillation occurrence on the whole night. The model allows to generate maps of predicted scintillation with antecedence from one to four hours. In summary, this thesis shows contributions with potential to create benefits on the scientific and technological scenarios in Brazil. Furthermore, monitoring data from the integrated database was made available to the scientific community via the software ISMR Query Tool, providing support to conduct new research in different institutions from Brazil and the world.
Ionospheric scintillations are rapid variation in amplitude and/or phase of a radio signal as it propagates through irregularities on electron density in the ionosphere. Such phenomenon degrades the performance of GNSS positioning, because it may cause accuracy degradation on observables and losses of lock, among other aspects. In Brazil, there is more susceptibility to occurrence of scintillations after sunset time between the spring and autumn equinoxes of years with high solar activity. Monitoring networks based on GNSS receivers are deployed over the Brazilian territory such as the CIGALA/CALIBRA network, managed by FCT/UNESP (with support from partners) in the last seven years. Monitoring data allows to develop several research regarding the scintillation effects, as in this thesis. Aspects regarding the modeling of ionospheric scintillation effects in Brazil were investigated, with emphasis on the mitigation of these effects on PPP and predictions of scintillation occurence. In the field of mitigation, existing approaches were investigated and a new one was proposed. The new approach for mitigation relies in both new functional and stochastic models for PPP, as well as a strategy to model effects of losses of lock. The proposed approach was tested with the scientific software RT-PPP and the achieved results were promising, including cases in which the expected accuracy for the PPP was recovered. In the field of the predictions, a database integration was conducted with data from three different networks (CIGALA/CALIBRA, ICEA e LISN). The integration allowed the development of a data-driven predictive model based on artificial neural network. The neural network is trained with data from previous nights. Data from the same night (around the sunset time) is also used to detect patterns regarding the emerging of the irregularities driving scintillation occurrence on the whole night. The model allows to generate maps of predicted scintillation with antecedence from one to four hours. In summary, this thesis shows contributions with potential to create benefits on the scientific and technological scenarios in Brazil. Furthermore, monitoring data from the integrated database was made available to the scientific community via the software ISMR Query Tool, providing support to conduct new research in different institutions from Brazil and the world.
Descrição
Palavras-chave
Cintilação ionosférica, GNSS, Mitigação, Predição, PPP, Ionospheric scintillation, Mitigation, Prediction