Research On Targets Information Fusion Of Radar And AIS Based On BP Neural Network

At present, with the rapid development of information and computer science, multi-sensorinformation fusion technology has been highlighted, especially in the area of military. For marinetraffic, radar and AIS are two important sensors to obtain the ship’s information, but the contentand accuracy of information from the two sensors are different, and they own respectiveadvantages and disadvantages, which can complement for each other. Therefore, it is quitenecessary and is actual significant to study on the information fusion of radar and AIS.First, this paper introduces some common knowledge on the application of multi-sensorinformation fusion technology and the neural network, particularly on the BP algorithm, basedon this, a model and research scheme of radar and AIS information fusion are proposed. Then,aimed at the kernel of the fusion, that is the targets association, a targets association algorithmusing BP network is put forward. And the important parts related to the design of networkstructure involved in the method are discussed in detail. After determining the associated targetswith this algorithm, at the stage of the radar and AIS information integration, the BP networkmodel is used again and the design of the very model is briefly describedFinally, large numbers of simulation and analysis about the referred algorithm and modelare carried out by Matlab and Excel. And the targets association algorithm is validated by theactual data obtained by the sensors furthermore. Besides, a comparative analysis on theintegration of radar and AIS information between the BP network and the traditionalweighted-statistic method is done. Simulation results corroborate that the proposed method iseffective and available.