Data fusion of GPS sensors using Particle Kalman Filter for ship dynamic positioning system

Depending on standards and class, dynamically positioned ships make use of different numbers of redundant sensors to determine current ship position. The paper presents a multi-sensor data fusion algorithm for the dynamic positioning system which allows it to record the proper signal from a number of sensors (GPS receivers). In the research, the Particle Kalman Filter with data fusion was used to estimate the position of the vessel. The presented algorithms generate a virtual measurement using three measurements from independent sensors. The performance of the Particle Kalman Filter algorithm was evaluated in simulation tests for two specific cases: in regular operation and when the signal of one sensor disappears.