Research Of Track Correlation Based On Multi-information

The purpose of multi-source information fusion is to process multi-sensor information effectively and to get more accurate and reliable results than a single sensor,resulting in the improvement of the entire system’s effectiveness.In recent years,multi-source information fusion is being developed rapidly in many fields,such as image fusion,artificial intelligence,remote sensing,and medicine.In multi-sensors information fusion,the randomness and uncertainty of measurement environment and sensors make the corresponding relationship of measurements fuzzy.To reestablish the corresponding relationship,correlation technology must be used.The result of track correlation will affect the accuracy and reasonableness of subsequent information processing directly.In the process of measurement,due to the measurement environment and the system itself,the generation of error is inevitable.To solve this problem,a method of track correlation based on spatial and temporal distribution is presented.Firstly,a distance function of tracks between two sensors is constructed by utilizing spatial and temporal distribution information.The function value is accumulated by time,and the cumulant is taken as the association statistics,based on which the nearest neighbor method is used to process the track association.Simulation results indicate that the proposed method is very efficient and robust to the system bias and random disturbance of the tracks.With the development of science and technology,sensors an measure more types of information of targets.The Track correlation is no longer limited to be based on the motion state information of targets.For the track correlation problem of dense targets or motion state information in a high degree of similarity,an association algorithm combined with the motion state information and the feature information is presented.A model of motion state information and feature information of track is established.The correlation of using each type of information used as associated statistics alone is analyzed.Simulation results indicate that combining the use of multiple types of information will result in a higher level of accuracy.Due to the difference of the ability of solving ambiguity,the contribution of each type of information to the correlation is different.Firstly,to judge the degree of merits of information,the dispersion,the magnitude of change and the error of each information is analyzed.Then,the degree of merits is quantified as the weight when calculating correlation and associated probabilities.The nearest neighbor method is used to process the track association.The simulation results indicate that the accuracy of information association is higher than only using motion state information of target.