Cognitive Tracking Waveform Design Based on Multiple Model Interaction and Measurement Information Fusion

Abstract

To enhance maneuvering target tracking in modern battlefield, cognitive radar could adjust its waveforms and information processing manner. In this paper, a novel adaptive waveform design method based on multiple model interaction and measurement information fusion is developed. First, some latest measurements and virtual ones are collected to exploit more robust information. Second, the unknown target state is formulated via the multi-model idea, and the tracking framework is highlighted by the matrix-weighted multi-model fusion (MMF) in lieu of the probability-weighted way. Finally, the MMF output covariance matrix is selected as the ellipse metric, and ellipse parameters can be obtained by using the eigenvalue decomposition. Given these parameters, fractional Fourier transform is used to rotate the measurement error-ellipse to make them orthogonal, and further obtain the desirable rotating orientations for the cognitive transmitting waveform. Simulations show that compared with several algorithms, e.g., MIMM and IMM, our algorithm could further improve tracking performance and robustness.

Keywords

• Cognitive waveform design
• eigenvalue decomposition
• multiple model interaction
• measurement information fusion
• fractional Fourier transform