Adaptive Strong Tracking Algorithm for Maneuvering Targets Based on Current Statistical Model

LIU Wang-sheng,LI Ya-an,CUI Lin
DOI: https://doi.org/10.3969/j.issn.1001-506x.2011.09.05
2011-01-01
Abstract:Combining the characteristic of the rise half normality fuzzy distribution function,an adaptive adjusting method of acceleration variance is presented based on Kalman filtering algorithm of current statistical model.It is composed of two sections of functions.The real maneuvering model is approached adaptively and the target is tracked accurately using this method.An adaptive adjustment means of maximum acceleration is given and the disadvantage of the model depending on maximum acceleration is overcomed.Tracking performance is enhanced for sudden maneuvering targets by introducing a strong track filter algorithm.The theoretical analysis and simulation results show that the match between maneuvering model and system mode is improved by using the algorithm.Performance for tracking strong maneuvering targets is enhanced and a good performance for tracking general motion is maintained.The algorithm is characterized by simple calculation,high tracking precision and easy realization.
What problem does this paper attempt to address?