An uncertainty-aware patent landscape of unmanned maritime systems: Search scope, priority-country portfolios, and technology convergence

Patent landscapes of complex systems are sensitive to search scope, family construction, geographic attribution, and classification. This study develops an uncertainty-aware landscape of unmanned maritime systems (UMS) using a Lens.org title-or-abstract search executed on 31 August 2026 and grouped by Simple Patent Family. The corpus contains 4448 families published in 2010–2025. Main inference uses 3271 families with earliest priority years 2010–2023; 625 families from 2024 form a right-truncation sensitivity cohort. A stratified model-based diagnostic classified 247 of 250 sampled families as relevant (98.8%; conditional 95% interval 96.5–99.6%), but it is not independent human ground truth. The analysis combines alternative growth curves, priority-country revealed technological advantage with bootstrap intervals, fractional CPC/IPC mapping, topic modeling, specification-robust co-classification networks, and assignee concentration. Bass has the lowest AIC, but model-dependent timing differs substantially, precluding a definitive lifecycle claim. Chinese-priority portfolios are relatively oriented toward computational intelligence, guidance/control, and communications, whereas US- and Korean-priority portfolios emphasize platform classes. B63G is the most stable bridge, and computational intelligence grows fastest from a small base. The findings support integration-oriented testbeds, interface-level portfolio monitoring, and reproducible search governance.