Tracking Filter Using Measurements of Target Extent with Intermittent Observation

Abstract

The tracking performance of traditional target tracking systems degenerates significantly when there is a drop of detection probability with intermittent observation. In this paper, a simple elliptical target model is proposed for exploiting sensor measurements of target extent, and then we design a filter for target tracking based on confidence weighted fusion with intermittent observation. Firstly, a measuring model by using measurements of target extent is built. Secondly, the Sequential Unscented Kalman Filter (SUKF) is presented based on the nonlinear measurements. Finally, for the four different cases of position and target extent detection channels, four sub-filters are designed respectively whose confidences are calculated based on the detection cases of the two channels, and then the output of the tracking filer is obtained by means of weighting the outputs of four sub-filters with the corresponding confidences. Monte-Carlo simulation results show that, with intermittent observation, the performance of the tracking system with measurements of target extent can be significantly improved as compared with that of the traditional system.