An Underwater Detection Method for UUV Intruder in Fluctuating Environment

Rapid advancements in Unmanned Underwater Vehicle (UUV) demand reliable early detection and warning systems. In this work, a detection method based on inevitable transmission variation was proposed. Based on the prior environmental conditions, fluctuated sound fields at different hypothetical target positions were calculated to pre-train partial least square model. A non-uniform sampling strategy was employed for target position selection to accommodate the detection difficulty along the source-receiver baseline. Following model training, multi-dimensional target characteristic space was constructed, and a non-linear fusion strategy was then implemented to enhance detection robustness. Simulation results demonstrated significant improvements in detection performance compared to conventional single-dimensional method. The proposed method provided a promising framework for UUV detection in complex underwater environments, offering enhanced reliability and adaptability.