LAUSR

Abstract Having a reliable and fast estimation of states regardless of prior knowledge about environmental conditions is a major goal to be achieved in application of Dynamic Positioning (DP). One way to have a faster estimation is to reduce state size which is studied in this research. The reduced elements belong to Wave Frequency (WF) model of estimation. Implemented estimation method is a new robust nonlinear extension of Kalman Filter called Square-Root Unscented Kalman Filter (SRUKF). Station-Keeping (SK) as one of the most common DP operations for supply vessels is considered as a case study in this research. Wind, wave and ocean current are considered as disturbing environmental loading. Computational efficiency and accuracy of 9 elements model is studied and compared with 15 elements counterpart in different sea Beaufort numbers (BN). Results show that in average 9 elements model is 37% faster than 15 elements model. However 15 elements model is generally more accurate, particularly in WF elements. The final results show that by assuming proper bounds for state Root-Mean Square Error (RMSE), using 9 elements model for BNs below 6 is a better choice.