Fuzzy-based Kalman Filter for ship navigation

For ship navigation system, the statistics of measurement noise will change with the actual working environment. If the traditional Kalman Filter is applied to integrated navigation system, the optimum state estimation may not be obtained even the filter becomes divergent due to the interference in environment. Based on fuzzy inference system (FIS), a new Kalman Filter was proposed in this paper. In the new filter, to make the theoretical value of measurement noise covariance close to its actual value, it was online regulated by using Fuzzy Inference System (FIS). Simulation results showed that the fuzzy-based filter performed well and the precision of the navigation system was improved.