Abstract With advancements in low-power control and high-energy-density batteries, unmanned underwater vehicles (UUVs) now achieve improved endurance for long-duration tasks in complex marine environments. However, multi-source time-varying uncertainties, such as time-varying ocean currents, biofouling, and subsystem degradation, accumulate over time, causing a decline in UUV performance during extended missions. To address this issue, this paper proposes a resilience-based design framework to ensure UUV performance throughout its entire lifetime. This framework treats uncertainty as an optimizable design variable, establishing a dynamic, closed-loop resilience assurance mechanism that spans the full “design–operation” lifetime. The framework is based on an uncertainty classification method that separates predictable uncertainties (described by evolutionary or distribution models) from unpredictable ones (characterized by model errors and limitations of current cognitive modeling). A two-phase resilience mechanism is established: 1) a preventive design optimization phase, where predictable uncertainties are incorporated into the design process to co-optimize UUV physical and control parameters for maximum longevity; and 2) an adaptive dynamic management phase, using online identification and adaptive control to compensate for unpredictable uncertainties in real time. Our approach ensures UUVs maintain performance by coordinating design and operation to dynamically handle evolving uncertainties. A case study of a long-endurance UUV, incorporating biofouling in the optimization phase and ocean currents in the management phase, demonstrates a 15% improvement in operational range and over 90% improvement in trajectory accuracy.
Yang et al. (2026) studied this question.