Optical Signatures and Detection Strategy for a Finned Bioinspired Unmanned Undersea Vehicle

The surface expressions for a submerged finned bioinspired unmanned underwater vehicle (UUV) were examined using VNIR and longwave infrared sensors as the system maneuvered in laboratory and field environments. Laboratory experiments revealed that the eddies generated by the flapping of the finned propulsion and attitude control system initially appeared as discrete thermal boils on the water surface. As these boils evolved, expanded, and merged into one another, two parallel thermal tracks were observed. The tracks cross-linked forming a single thermal swath or footprint behind the trajectory of the vehicle. Similar thermal disruptions were observed when experiments were performed in an uncontrolled harbor environment under daytime and nighttime lighting conditions. Estimates for the background clutter, system signal, and detection statistic were generated using probabilistic models to demonstrate the feasibility of extracting signals from complex environments in both laboratory and harbor experiments. Performance of the detection model was presented in the form of receiver operating characteristic curves. Results clearly demonstrate that we can achieve > 99% probability of detecting the presence of the UUV with a very low probability of false alarm (<; 0.005%) in a real harbor environment, where we estimate $\text{signal-to-noise ratio}=21.5$.