Intelligent Multisource Autonomous Navigation: Review and Perspectives

Multisource navigation for autonomous platforms has progressed rapidly, yet resilience under sensor faults, model error, and complex disturbances remains limited. This article surveys recent advances across sensor perception, state estimation/fusion, and system-level assessment, and identifies recurring limitations including coupled non-Gaussian/nonindependent and identically distributed (non-IID) disturbances and model uncertainty. The Intelligent Multisource Autonomous Navigation (IMAN) framework is proposed, which models navigation as hierarchical information processing and defines four system characteristics (detectability, reconfigurability, dependability, and completeness). The framework aims to provide a common analytical language for designing and evaluating unified, deployable multisource navigation systems. Representative methods are mapped to the IMAN characteristics to highlight areas of substantive progress and remaining gaps. Looking forward, priority directions are the development of formal disturbance tools and resilience assessment standards.