Netted Radar Fusion Algorithm Based on Converted Measurement Reconstruction

Multi-radar network detection is an effective approach to deal with stealth and low altitude penetration. Converted measurement Kalman filter (CMKF) is widely used in netted radar fusion tracking due to its simple implementation and low computational complexity. The conversion measurement of CMKF is a pseudo linear expression of the target state containing strong nonlinear components, which can cause a decrease in fusion accuracy especially in the case of large azimuth measurement errors. In order to suppress the influence of nonlinear errors, theoretical analysis is conducted on the error influence of classical converted measurement models, and it is concluded that the cosine of azimuth error is the main factor causing nonlinear errors. Based on this argument, a more precise converted measurement model is reconstructed with the predicted azimuth. By adaptively correcting the azimuth measurement during the filtering process, the nonlinear influence of the cosine of the orientation error is effectively suppressed. Theoretical analysis and simulation results show that the improved CMKF applied to the netted radar system significantly reduces nonlinear errors and improves fusion accuracy.