Application of the Cramer-Rao lower bound for bearing estimation to STAP performance studies

• Published in: Proceedings of the 2004 IEEE Radar Conference (IEEE Cat. No.04CH37509) pp. 402 - 407
• Main Authors: Showman, G.A., Melvin, W.L., Zywicki, D.J.
• Format: Conference Proceeding
• Language: English
• Published: Piscataway, NJ IEEE 2004
• Subjects: Antennas; Applied sciences; Detection, estimation, filtering, equalization, prediction; Direction of arrival estimation; Doppler radar; Exact sciences and technology; Geometry; Information, signal and communications theory; Maximum likelihood estimation; Organization and planning of networks (techniques and equipments); Parameter estimation; Radar antennas; Radar detection; Radiocommunications; Radiolocalization and radionavigation; Sensor arrays; Signal and communications theory; Signal to noise ratio; Signal, noise; Spaceborne radar; Systems, networks and services of telecommunications; Telecommunications; Telecommunications and information theory; Performance evaluation; Direction-of-arrival estimation; Parameter estimation; Signal estimation; Network architecture; Algorithm; Implementation; Accuracy; Space-time adaptive processing; Radar; Maximum likelihood; Antenna array; Cramer Rao inequality
• ISBN: 078038234X; 9780780382343
• DOI: 10.1109/NRC.2004.1316458

The performance of space-time adaptive processing (STAP) radar is a strong function of array geometry and the particular algorithm implementation. Traditionally, detection performance has been of paramount importance, but recently interest has grown in the accuracy of target direction of arrival (DOA) estimates. This paper describes an evaluation of the Cramer-Rao lower bound (CRLB) for DOA accuracy. The CRLB is useful for bounding the bearing estimation performance of candidate array architectures and STAP algorithms, but often generates counter-intuitive results. Anomalous characteristics of the CRLB are investigated, explained, and placed in the context of maximum likelihood estimation (MLE) properties. The end product is a tool that can be applied to comparative analyses with confidence. The paper concludes by demonstrating the utility of the CRLB in both array and algorithm studies.