An Dynamic Kalman Filter Restraining Outliers For Target Tracking

During Kalman filter applies the outliers included in observation values are important factor which affect filtering performance.The paper analyses the mechanism that outliers affect Kalman filter,and presents a set of methods distinguishing and dealing with outliers according to arithmetic's principle and engineering practices.Simulation calculation demonstrate that the set of methods is characterized by high reliability and simpleness.It can eliminate the bad effect which outliers affect filter,and improve the filter accuracy.