Advanced Ship Motion and Air Wake Sensing and Prediction

"Landing on a carrier that is moving about in rough seas is a difficult piloting skill, whether flying fixed wing aircraft, helicopters, LCAC or other assault craft. Many attempts have been made over the past 30 years to predict the ship motions at least 10to 20 seconds into the future so the pilot could know where the `target' would be when he got there. A less ambitious goal is to tell the pilot when the ship will be in a relatively quiescent state. Most of the work has been purely statistical, where`optimum' predictors of various forms were used to estimate the current ship dynamic state, and propagate that state based on the recent time history and stochastic input. Neural nets, Kalman filters, fuzzy logic and (last year) even chaos theory has beentried, but none with demonstrable success.Our approach is to instead combine deterministic and statistical prediction methods. New low cost instrumentation, including DGPS and MEMS inertial sensors are used by small expendable ROVs to `preview' the incoming wave field. The surface elevationprofile is deterministically projected forward in space/time using Fourier transforms and the dispersion relation of deep-water ocean waves. A linear strip theory ship model continually restarts with the current state estimate and propagates forward usingthe anticipated waves as forcing function, providing improved estimates of future ship motion. In addition to its military uses, motion prediction