Navigating the Future: A Kalman Filter-Based Global-Local-Global (GLG) Position Estimation for Autonomous Vehicles in GPS Denied Environments

Accurate navigation in Global Positioning System (GPS) denied environments, such as tunnels, rural areas and underwater regions is a critical challenge for autonomous vehicles (AVs). Traditional position estimation methods face challenges due to their computational complexity and inability to consider external factors like sensor drift and human input errors. This paper introduces a Global-Local-Global (GLG) position estimation method that integrates advanced inertial sensors with navigation sensors, utilizing Kalman Filter (KF) and Extended Kalman Filter (EKF) to enhance trajectory accuracy. The proposed mechanism is validated through analysis and real-time tests using custom-built hardware in terrestrial and underwater environments. It demonstrates effectiveness during GPS outages lasting 5–60 seconds, achieving position drifts of 2.84 m for GLG, 1.35 m for GLG + KF, and 1.27 m for GLG + EKF over a total distance travelled of 1.1 km. This method is particularly suited for short-range applications, including aqua farming, tunnel navigation, and dense areas, providing a robust solution for accurate position estimation in challenging scenarios.