Research On Data Processing Technology Of Multiple Platforms Shipborne Radar System

In recent years, as the huge economic benefit of the marine resource widelyaccepted by the world, more and more conflicts had happened between countries.Under this situation, it was of great significance to set up a powerful and effectivecoast guard and monitoring system to track ships near the exclusive economic zoneof China. The paper was aimed at researching data processing technology of themultip le platforms radar system. Through useful processing procedure, targets couldbe fixed and tracked. Then more accuracy and effective motion information wasprovided to the sea surveillance system.Data pre-processing technologies were introduced first of all, includ ing spaceregistration and time registration algorithms. Space registration algorithm turneddatum from different coordinates into the same one. Time registration algorithmconverted high sampling rate to the lowest rate to ensure all the radars in the systemhad the same rate. Simulation results analyzed the impact of registration algorithmson data accuracy. By introducing Joint Probabilistic Data Association (JPDA)Algorithm and Interacting Multiple Model (IMM) Algorithm, IMM-JPDA wasproposed based on the two algorithms. It was used for tracking motoring targets indense clutter environment. In the IMM-JPDA algorithm, every filter was replacedby JPDA filter. The superiority of the combination was that it could realize not onlydata association but also motoring targets tracking. The algorithm had highreliability and was very practical, but with high computational complexity and badproperty with real-time. For non-monitoring targets, tracking accuracy didn’timprove at all. In the fusion center, track association and track fusion algorithmswere stated. Track association algorithms based on fuzzy mathematics wereintroduced in detail. They had good properties when the number of radars was onlytwo. But when the number became greater, they had to face MultidimensionalAllocation Problem (MAP). Track association algorithm based on Hopfield NeuralNetwork was adopted then for solving the MAP. Detailed derivation aboutTwo-dimensional Assignment Problem and Multid imensional Allocation Problemwas given. Much more discussion was given about parameter setting rules.Simulation clearly demonstrated the procedure of track association. For theassociated tracks, dynamic weighted fus ion (DWF) algorithm based on fuzzymathematics and simple fusion algorithm (SF) were employed for track fusion.DWF algorithm imported fuzzy means clustering algorithm to calculate the dynamicweights of associated tracks. SF calculated the targets’ state covariance as fusion weights. Numerical simulations were conducted to compare properties of the twoalgorithms from the aspects of accuracy, reliability and validity. Graphical UserInterfaces (GUI) was designed for simp le human-computer interaction and revealedthe simulation results stratified at the end.This paper made an intensive research on data processing technology ofmultip le platform shipborne radar system and obtained some valueable results.Much deeper research is needed for the improvement of the system’s shortage.