Search

In rotating radars, the revisit interval is time variant and uncertain due to its dependence on target states and the disturbance of process noise, violating the assumption of prior known revisit interval in conventional tracking algorithms. In this article, the formulation of state transition and target state estimation with revisit interval uncertainty are investigated. The dependence of the revisit interval on the antenna rotation and target states is formulated to describe target dynamics, together with the traditional state transition equation. A space–time joint processing method based on Newton–Raphson iteration is presented to calculate the revisit interval and the predicted state simultaneously. Such a numerical solution is considered as a general function reflecting the target state transition. The unscented Kalman filter (UKF) is then utilized to deal with the nonlinear filtering problem and produce state estimates. In addition, a filter initialization method based on the target dynamics and the unscented transformation is presented to produce consistent initial state estimates. Experimental results illustrate the superiority of the proposed space–time joint processing unscented Kalman filter (STJP-UKF). The performance of the STJP-UKF is superior to that of the standard UKF, and comparable with that of the UKF with known revisit intervals.

In tracking applications without angle measurements, multiple sensors are usually applied. Accurate range-Doppler (R-D) estimates with a single sensor play an important role in clutter eliminating and date fusion in multisensor target tracking. In this article, maneuvering models are first established in the range-squared (RS) coordinate, then the R-D estimation of a maneuvering target is investigated. According to the Cartesian nearly constant acceleration and constant turn models, pseudostate vectors are defined in the RS coordinate and linear time-evolution equations of pseudostate are derived to describe the corresponding maneuvers. The motion models are incorporated into the converted measurement Kalman filter to extract R-D estimates from converted measurements. Corresponding to the models, the filter initialization approach, which utilizes the prior information on motion model and kinematic parameters of target, is developed. Based on the model set in the RS coordinate and the interacting multiple model (IMM) method, a recursive estimator is developed to produce pseudostate estimates of a maneuvering target. Final R-D estimates are provided through the conversion adopting the scaled unscented transformation outside the estimation recursions to avoid the propagation of nonlinearity approximation errors. Numerical experiments in different scenarios demonstrate that the maneuvering models accurately describe the time-evolution of pseudostate and the IMM estimator achieves superior estimation accuracy and consistency thanks to the exploitation of linear motion models.

In multisensor tracking systems, bias compensation is essential for accurate data fusion. It is challenging to obtain correct fusion results in the presence of both spatial and temporal biases, as well as target maneuver. In this article, a sequential spatiotemporal bias compensation and data fusion method is proposed for maneuvering target tracking with asynchronous multisensor measurements. The spatiotemporal biases of the sensors are augmented into the state vector to be estimated. The state equations are presented to formulate the nearly coordinated turn motion and the nearly constant acceleration motion, respectively, without exactly known measurement interval. In each target motion, the measurements are formulated as functions of both spatiotemporal bias and target state based on the time difference between the measurements and the states to be estimated. Furthermore, the interacting multiple model estimator is incorporated with the unscented Kalman filter to achieve simultaneous sequential estimation of spatiotemporal biases and target states in the presence of target maneuvers. Finally, the posterior Cramer–Rao lower bound for spatiotemporal bias and state estimation is provided. Simulation experiments are performed to demonstrate the effectiveness of the proposed method.