Banca de DEFESA: EDUARDO AUGUSTO DE OLIVEIRA BORGES

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
DISCENTE: EDUARDO AUGUSTO DE OLIVEIRA BORGES
DATA: 09/02/2026
HORA: 07:00
LOCAL: meet.google.com/hmr-qoxf-hes
TÍTULO:

Implementation of Precise Point Positioning GPS-aided Inertial Navigation Algorithms in Real Time Embedded Systems

PALAVRAS-CHAVES:

Navigation System. INS/GNSS Integration. RT-PPP.

PÁGINAS: 80
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
SUBÁREA: Medidas Elétricas, Magnéticas e Eletrônicas; Instrumentação
ESPECIALIDADE: Instrumentação Eletrônica
RESUMO:

This dissertation presents an integrated approach to enhancing positioning accuracy and system robustness in connected agricultural vehicles through the fusion of Inertial Navigation Systems (INS) and Global Navigation Satellite Systems (GNSS), with emphasis on the implementation of Real-Time Precise Point Positioning (RT-PPP) technology using low-cost single-frequency receivers. The study systematically evaluates various integration techniques, focusing on mitigating inherent errors from both systems—INS drift accumulation and GNSS signal degradation—through optimized Extended Kalman Filter (EKF) covariance matrix settings and the development of real-time positioning algorithms comparing Loosely Coupled (LC) and Tightly Coupled (TC) integration methods. Three major contributions to the field have been realized through peer-reviewed conference articles. The first article demonstrates that LC INS/GNSS integration with automotive-grade sensors achieves 31%, 23%, and 22% improvements in horizontal, vertical, and total positioning accuracy over standalone GNSS. The second article analyzes EKF measurement covariance optimization across RT-PPP-GNSS-IWLS, RT-PPP-GNSSEKF, LC-INS-RT-PPP-GNSS, and TC-INS-RT-PPP-GNSS algorithms using IGS corrections, revealing TC integration delivers <2 m horizontal accuracy even with single-satellite visibility via proper pseudorange uncertainty tuning. The third article presents the real-time C++ implementation and urban vehicle validation of PPP-based TC INS/GNSS on ARM64 architecture, achieving micrometer-level numerical consistency with post-processed reference solutions while maintaining deterministic execution under GNSS-challenged conditions. Results were validated through extensive experimental trials with real vehicle data across rural, urban, and mixed environments, confirming the effectiveness, robustness, and computational feasibility of the proposed methodologies for precision agriculture applications requiring continuous sub-meter navigation under heterogeneous satellite visibility and dynamic maneuvers.

MEMBROS DA BANCA:
Externo à Instituição - CARLOS RENATO CAPUTO DURÃO - UFRJ (Membro)
Interno - DANILO ALVES DE LIMA (Suplente)
Externo à Instituição - VINÍCIUS MANHÃES GABRIEL DE BRITO CAVALCANTI - IFF (Suplente)
Presidente - FELIPE OLIVEIRA E SILVA (Membro)
Interno - DEMOSTENES ZEGARRA RODRIGUEZ (Membro)