Research On Data Fusion Technology For Wireless Sensor Network In Ship Dynamic Positioning System

As the range of marine for people to develop and explora is gradually expanding, for the ships of Marine operation, the Dynamic Positioning Control System has become one of a core component for the ship. The Dynamic Positioning Control System of the Marine has a variety of position reference systems and sensors to monitor the environmental factors, the status of ship and the conditions of ocean, for which we must adopt Data Fusion for the measurements in order to improve the accuracy and reliability of the data and System Fault Tolerance. In this article the distributed data processing structure of the emerging Wireless Sensor Network (WSN) and the centralized-bus data processing structure are combined, the Data Fusion method based on Wireless Sensor Network for the Dynamic Positioning Control System on ship is researched.Firstly, for the characteristics of monitoring function on data fusion technology in Dynamic Positioning Control System on ship.The fusion method based on the collaboration between the nodes is adopted, the dynamic group clusters can be determined by the qualitative analysis of sensor nodes; the optimization method based on Mobile Agent and Ant Colony Traverse for cluster data fusion is proposed and the fuzzy adaptive Kalman filter method based on the clustering structure for heterogeneous multi-sensor fusion is researched.Secondly, for the characteristics of decision-making and control-oriented functions on the data integration in Dynamic Positioning Control System on ship, the hard integration model of the FPGA-BP is built by making full use of the advantage of the hardware resources of the parallel computing, which make such complex prediction algorithm of BP neural network model based on the nodes can be achieved in WSN network; The Support Vector Machine prediction algorithm based on disjoint collection is proposed.This model has two advantages:firstly, the use of strong systolic array-based BP neural network fusion method in aggregation node can real-time optimiz Support Vector Machine model to overcome the slow training speed, falling into local optimum easily and over learning problems, it can ensure the prediction accuracy; On the other hand, Support Vector Machine prediction calculation function can be dispersed to the DJC aggregation nodes by setting DJC model in WSN network, and each DJC aggregation node can selectively upload data, so that not only the computational burden of the aggregation node is reduced,but also can effectively implement complex optimal control algorithms.Finally, how to use the WSN Data Fusion method in ship Dynamic Position System is researched, the distributed control structures is built, the Fuzzy adaptive Kalman filter algorithm used in estimating the status of ship is researched, The Quadratic Optimal Control algorithm performance in the Ship Dynamic Positioning MPC can be solved by Support Vector Machine prediction algorithm based on the DJC model in Wireless Wensore Network, getting the optimal thrust and results. For the characteristics of network and transmission functions in data fusion technology, the data fusion routing algorithm based on sleep scheduling mechanism is studied, data fusion method whether in ordinary sensor nodes or sink node are based on reasonable and reliable data fusion structure, enhancing the ability to avoide entire network congestion, to reduce the delay of the data transmission and to improve the real-time nature.Overall, this paper mainly around the clustering topology of wireless sensor networks for the Dynamic Positioning System on ship, Data Fusion technology in sensor nodes is researched from the data level, feature-level and decision-making level. The results of research will greatly increase reliability and real-time of data acquisition and the accuracy of the predictive control in the Ship Dynamic Positioning Control System.