Research On Vessel Position Reference System Asynchronous Information Fusion Method

In modern ship dynamic positioning system, in order to improve the accuracy and reliability of the ship position information, the position reference systems were redundant configuration. Redundant configuration of homogeneous and heterogeneous position reference system can provide a lot of ship movement information, how to make the use of redundant information and ensure the accuracy and reliability of the information obtained in the dynamic positioning system is the core issue to be resolved.According to the ship kinematic and dynamic positioning vessel movement characteristics, a dynamic positioning vessel kinematic model was established, then on basis of the measuring and positioning principle of ship position reference system, the paper presented a measurement model of ship position measuring system,both the kinematic model and the measurement model were prepared for the subsequent chapters.Combining with multi-sensor information fusion structure, the position reference system information fusion included two parts: position reference system measurement data processing and information fusion algorithm, wherein the measurement data preprocessing consist of measurement information space registration, data quality detection and filtering.For the quality detect of the sensors measurement data was consistent with subsystem fault detection, accordingly fault detection algorithms was build to detect the quality of measurement data and the subsystems fault, and the fault detection algorithm included mutant fault detection method and graded detection method; and UKF filtering algorithm was choosen to estimate the state of the ship position.In practical applications, there are two asynchronous sampling information fusion approaches, one is time registration, and the other one is voting logic. But both of the two methods have significant deficiencies, such as the poor real-time of time registration and the voting methods fail when common mode failures exist in the system, thus the fusion accuracy of the system will be restricted by the deficiencies. However, the fusion algorithms included synchronization and asynchronous data fusion algorithm, the article put forward the synchronization multi-sensor data optimal distributed estimation fusion algorithm; and in allusion to asynchronous sampling information, the asynchronous information fusion algorithm based on the scale was proposed, which made the entire system built on different scales, the approach achieved the direct integration fusion of asynchronous sample information, therefore the fusion performance was improved.According to the actual needs of the redundant dynamic positioning system and fault detection algorithm combined with the established multi-sensor synchronous and asynchronous information fusion algorithm, the synchronous and asynchronous information fusion structure and fusion algorithm then were designed. Finally, the simulation results demonstrated the effectiveness of the position reference system asynchronism information fusion structure and fusion algorithm.