Autonomous Navigation Systems in GPS-Denied Environments: A Review of Techniques and Applications

Autonomous navigation entails using systems programmed to navigate predefined channels without human intervention. From driverless vehicles to RPAS (remotely piloted aircraft systems) and UAVs (unmanned aerial vehicles), the introduction of autonomous navigation systems has revolutionised logistics, healthcare, aerospace, transportation, and numerous other industries. Traditionally, autonomous systems have relied on GPS (Global Positioning System) for accurate navigation and positioning with wide application in various autonomous systems, including precision-guided missiles, logistics and fleet management, and turn-by-turn navigation in driverless vehicles. While the introduction of GPS improved autonomous navigation significantly, it is not sufficiently reliable for fully autonomous navigation due to a reliance on external signals which are typically vulnerable to signal jamming and obstruction, interference, signal loss, and multipath errors. This limits the operations of autonomous navigation systems (ANSs) in GPS-denied environments where satellite or external signals are degraded, jammed, reflected, or obstructed. Consequently, environments without GPS availability render it impossible for ANSs to perform localisation swiftly and robustly while avoiding obstacles with no priori information of the location. This therefore also limits the systems' computation and perceptive capability. As such, applications and techniques that introduce disruptive technologies, such as machine learning, deep learning, artificial intelligence, and sensor technology, have been developed to overcome some of these inherent limitations of GPS. This paper evaluates how the integration of such techniques facilitate ANSs in determining their relative positions in GPS-denied locations, ensuring operational efficiency and effective navigation. A combination of systematic literature reviews, descriptive analysis, and comparative analysis research methods is used to facilitate a comprehensive analysis and synthesis of relevant work and techniques used to overcome GPS-denied settings, as well as a comparison of these various techniques and their applications based on specific criteria-such as usability, efficiency, performance, and limitations-to identify the weaknesses and strengths of each method.