Statistical modeling of consecutive range profiles for radar target recognition

The high resolution range profile (HRRP) is known as the most important tool in radar target recognition. Mainly, the sensitivity to the aspect angle limits the performance of the related methods. To overcome this problem, Gaussian mixture distribution is used to model the short-term relation of consecutive HRRPs. In this work, an alternative dynamical system based method is proposed to overcome the limitations of recent methods in the field such as the independency assumption. Here, the performance of the method is tested by the data produced in an electromagnetic simulation for the radar return from an aerial maneuvering target. The results show the performance of the proposed method comparing to Gaussian mixture and factor analysis based methods using the Akaike information criterion (AIC).