<p>Precise positioning is essential for navigation systems, particularly in scenarios where GPS signals are compromised, requiring alternative sensors to ensure continued accuracy. This paper presents a robust multi-sensor fusion system that integrates GPS, an Inertial Navigation System (INS), and Visual Odometry (VO) to achieve high-precision localization, even in challenging environments. The proposed system employs a shallow Convolutional Neural Network (CNN) and Kalman Filter (KF) derivatives, providing a power-efficient alternative to more complex and resource-intensive architectures, such as CNN-RNN combinations. The system effectively addresses the dual challenges of GPS outages and visual distortions by leveraging these lightweight components. During GPS signal interruptions, the system generates pseudo-GPS outputs via a neural network, which are integrated with inertial and visual data and processed through advanced algorithms to maintain accurate positioning. Furthermore, the system is designed to withstand severe visual distortions caused by remote laser attacks, ensuring reliable performance. Simulation results demonstrate that under harsh conditions, including complete GPS outages and external disruptions to the visual component, the proposed GPS/VIO system outperforms standalone INS and highly robust Visual-Inertial Odometry (VIO) systems by 82.82% and 74.25%, respectively, in terms of accuracy.</p>