Search

By utilizing the complex monopulse ratio (CMR) computed from the digital phased array (DPA) radar, a method for estimating the parameter of retrodirective cross-eye jamming (RCJ) is proposed, which is named as CMR-DPA. In CMR-DPA, the sum and difference beam outputs of an RCJ signal in DPA for a given receive beam direction are derived. CMR induced by RCJ is calculated and found to be determined by the receive beam direction and RCJ parameters. Therefore, a cluster of receive beam directions is appropriately selected and multiple CMRs are measured. Then the RCJ parameter estimation criterion is formulated based on the CMR results. A genetic algorithm is employed to solve the estimation problem. The simulation results demonstrate that CMR-DPA can estimate the RCJ parameters with high accuracy.

For networked radars on moving platforms, registration errors caused by time-varying relative positions and orientation angles deteriorate the target tracking accuracy significantly. In this article, a spatio–temporal alignment method is proposed for networked radars on moving platforms by sharing the only parameter of coherent processing intervals. The time-varying spatio–temporal alignment parameters of moving radars are modeled as linear expressions of discrete cosine transform basis functions. The proposed method does not require the knowledge of positions and orientation angles of moving radars at each sample time and is suitable for asynchronous measurements from networked radars. For the registration objective function, the alignment parameter estimation using the steepest descent method is derived. In order to solve the problem of local optimal solution caused by the nonconvexity of the objective function, a filling matrix-based method for solving the global optimal solution of the registration function is proposed. Furthermore, the Cramér–Rao lower bound of the alignment parameter is derived for moving radars, which is compared with estimation errors of parameters. Simulation results show that this proposed alignment algorithm for moving radars improves the target tracking performance compared with the existing methods.

Real-time calibration of time-varying sensor position errors for flexible arrays is crucial to alleviate the array performance deterioration. Previous studies shown that sensor arrays of any shape can be self-calibrated except for the nominal linear array with small sensor position errors. In this article, the self-calibration restriction of the nominal linear array is resolved, and an improved self-calibration algorithm is proposed to jointly estimate source direction of arrival (DOAs) and sensor positions. First, the identifiability of unknown parameters including sensor positions and source DOAs in array receiving model is analyzed. It is found that the failure of self-calibration for the nominal linear array is caused by the ambiguity of sensor position estimations under unknown source DOAs. Next, the accurate priori distance of one sensor pair in the array is introduced to solve such ambiguity. Moreover, the hybrid Cramér–Rao bound is also derived to verify the identifiability of unknown parameters under such distance priori. Finally, an improved sensor position self-calibration algorithm is proposed, which has lower computational complexity compared with the state-of-the-art self-calibration method. Numerical results demonstrate the self-calibration capability of the nominal linear array with small positional errors under the distance priori and the effectiveness of the proposed algorithm.

Model set design is an important area for multiple-mode estimation, and its main purpose is to design a suitable model set to improve the accuracy and stability of target tacking. In this article, a novel multiple model estimation algorithm, namely, centroid fixed structure of interacting multiple-model estimation (CFIMM), is proposed to obtain the characteristic of the high precision and strong stability. First, the minimum distance method and the minimum model set method are, respectively, provided. Then, based on those two methods, the centroid model set design method is proposed with three different approaches to split the centroid. It proves that the centroid model set has minimal mathematical expectations with the actual models, when the unknown real model space is very large and even uncountable. Finally, the processing of the proposed CFIMM algorithm based on the centroid model set design method is discussed in detail. The proposed CFIMM holds not only the advantages of centroid model set but also the characteristic of the high precision and strong stability. The simulations of CFIMM highlight the correctness and effectiveness of the proposed methods.