Research On Target Tracking Algorithms With Constraints

Both locating and tracking are the basic needs in the field of wireless sensor network and navigation. In a real environment, there are a variety of factors affecting the accuracy of locating and tracking. These factors collectively referred to as constraint information. In the different types of constraint information, some of them to be used to improve the accuracy of locating and tracking which are referred to as prior knowledge. Some constraint information will reduce the accuracy of locating and tracking. This kind of constraint information can be corrected by studying the law of the phenomenon.In this paper, two kinds of constraint information are discussed. The first kind is the quadratic equality constraints of the target that has a beneficial effect of improving the accuracy and can be used as prior knowledge. This corresponds to the scene of single observer passive tracking problem with quadratic constraints of the target. The second kind is the non-line-of-sight error, which has an unfavorable result of the accuracy. The impact can be reduced by mitigating the NLOS error. The NLOS corresponds to the multi-station positioning and tracking in the NLOS environment.In this paper, the locating and tracking algorithms with two kind of constraint information are discussed in the following aspects. By comparing with the existing methods, several algorithms are proposed with more superiority.1) For a single observer passive tracking problem with quadratic constraints of the target,a target-tracking algorithm based on cubature Kalman filter with projected quadratic equality constraints is proposed in this paper. The algorithm is based on the combination of the constraint projection theory and the cubature Kalman filter. Simulation results show that the proposed algorithm outperforms previously developed algorithms with quadratic equality constraints in terms of enhanced target tracking accuracy.2) For a single observer passive tracking problem with quadratic constraints of the target,a target-tracking algorithm based on cubature Kalman filter with truncated quadratic equality constraints is proposed in this paper. The algorithm is based on the combination of the probability density truncated theory and the cubature Kalman filter. Simulation results show that the proposed algorithm outperforms previously developed algorithms with quadratic equality constraints in terms of enhanced target tracking accuracy.3) Based on the probability density truncated theory, extending the quadratic constraints to quadratic inequality constraints, a target-tracking algorithm based on cubature Kalman filter with truncated quadratic inequality constraints is proposed in this paper. Simulation results show that the proposed algorithm enhances the target tracking accuracy in this scene.4) For the multi-station positioning and tracking in the NLOS environment, based on the on the disk of scatterers model theory, a NLOS mitigation location method based on scatterers model is proposed. Simulation results show that the proposed algorithm outperforms previously developed algorithms in terms of enhanced target positioning accuracy.5) Based on the on the disk of scatterers model theory, combined with cubature Kalman filter, a target-tracking algorithm based on scatterers model theory and cubature Kalman filter in the NLOS environment is proposed in this paper. Simulation results show that the proposed algorithm efficiently mitigating the NLOS error, improving the tracking accuracy.