Скачать или смотреть Ashwin Kanhere Defense: Resilient GPS Positioning using DNNs and Sensor Fusion with FGO

Civilian Global Positioning System (GPS) forms the bedrock of position, velocity, and timing (PVT) estimation for maritime, aviation, search and rescue, and timing applications across the globe. However, civilian GPS signals and these applications are vulnerable to spoofing and measurement faults.

In this presentation, we discuss methods to provide resilient positioning by mitigating the effects of GPS spoofing and faults. First, we propose a method for improving GPS positioning in the presence of faults using corrections from deep neural networks. Second, we propose a method for spoofing resilient GPS positioning that uses upcoming signal enhancements and odometry sensors in a factor graph. Third, we propose a method for joint mitigation of GPS spoofing and measurement faults using an expectation-maximization-based approach. These proposed approaches provide avenues for GPS positioning that increase resilience to these GPS vulnerabilities.