Radar Array Processing of Experimental Data Via the Scan-MUSIC Algorithm

Report No. ARL-TR-3135
Authors: Canh Ly
Date/Pages: June 2004; 57 pages
Abstract: This report demonstrates a novel algorithm of radar array processing that was developed at the U.S. Army Research Laboratory (ARL) and named the Scan-MUltiple SIgnal Classification (SMUSIC) algorithm. The algorithm can improve angular resolution of real targets when a single rotatable antenna is used as the sensor. The algorithm is an adaptation and extension of the MUSIC algorithm which requires an array of many sensors. We achieved a resolution of approximately 1/3 of antenna beamwidth for all target orientations and range offsets from experimental data. Previous work that used closely spaced point targets showed that the results from the SMUSIC algorithm could be degraded because of strong interference between the point target returns. Those interferences occur because of the relative phases of their returned signals. Therefore, there were some limitations of the algorithm for simple targets that consisted of a single scatterer. This new work shows that the limitation does not occur for real targets, which are more complex, such as would be present in Army radar applications. We used a Ka-band, stepped frequency, pulsed radar to acquire the returns from two real targets, both located within a main lobe of the antenna pattern. The SMUSIC algorithm, which used magnitude-only experimental data, was then applied to accurately locate the centroids of the targets (an M60 tank and an M113 armored personnel carrier). All routines needed are written in MATLAB® and are included in appendix A. MATLAB® is a registered trademark of The MathWorks.
Distribution: Approved for public release
  Download Report ( 0.734 MBytes )
If you are visually impaired or need a physical copy of this report, please visit and contact DTIC.

Last Update / Reviewed: June 1, 2004