Interacting multiple sensor unscented Kalman filter

Due to the log-normal model of the received signal strength(RSS), the range measurements have variance proportional to their actual range, and so this results in degradation of the tracking performance with the range increasing. To deal with this problem, we consider the collaborative tracking procedure in a cluster as a Markov jump nonlinear system, and the design the interacting multiple sensor unscented Kalman filter(IMSUKF) algorithm via multiple measurement models in a cluster, which is different with the interacting multiple model(IMM) algorithm. This approach consists of three parts: one-step unscented Kalman filter sensor, probability update, and estimate fusion. Finally, simulation results show the effectiveness of the proposed method.