摘要：In radar or sonar systems, we obtain measurements of kinematic information (range, bearing and velocity) as well as non-kinematic information (amplitude) for signal processing. Considering nonkinematic information, data association process can be improved when the kinematic information is not enough to obtain the reasonable performance in severely cluttered environment. The tracker exploits nonkinematic information in the form of signal to noise ratio (SNR). That means that to incorporate the nonkinematic information we need to know the target SNR that is not known in advance. To overcome the restriction, we introduce the marginalized SNR. Moreover, we consider nonlinear estimation problem in this paper. So, the performance degradation from nonlinearity in systems should be alleviated. We propose to apply particle filter with maximum probability data association strategy for target tracking in nonlinear dynamic systems using both kinematic and non-kinematic information. Simulation results demonstrate the effectiveness and high estimation accuracy of the idea of combining particle filtering, marginalization of unknown SNR and maximum probability data association strategy.
关键词：Particle filtering; Amplitude feature; Probability data association
