A novel overall adaptive algorithm for maneuvering target tracking

An adaptive tracking filter is proposed to enhance the tracking performance of maneuvering target. After analyzing the drawback of input estimation, an assumption that acceleration over detection window is linear-distributed is made. The assumption greatly looses the constant input constraint and makes input estimation with long detection window size applicable. The filter modal of input estimation based on linear acceleration assumption is deduced. Combined with innovation modification technique, the modified algorithm is more robust and adaptive. Simulations results verify that the modified IE performs better than original IE in tracking of maneuvering target with variable acceleration.