Performance Comparison of Cell Averaging and "greatest-of" Constant False Alarm Rate (CFAR) Methods

Résumé du livre

Radar system detection procedures involve the comparison of a received signal to a threshold value. A fixed threshold is not successful in controlling the false alarms because of changing interference conditions such as noise, clutter, and jammers. Consequently, an adaptive threshold is frequently used to establish a constant false alarm rate (CFAR). The maximum likelihood cell averaging approach is one of the adaptive procedures. The cell averaging CFAR is of major importance because it has superior performance in homogeneous Gaussian noise. Unfortunately, nonhomogeneous interference such as that produced at chaff and clutter edges will degrade the performance of the cell averaging CFAR. The search for a CFAR method which can control the false alarms resulting from clutter edges has produced a modification to the cell averaging approach which is known as the 'greatest-of' CFAR. This effort compared the performance of the cell averaging and 'greatest-of' CFAR methods for different processor configurations. The performance comparison was made based on probabilities of false alarm and probabilities of detection determined by a Monte Carlo simulation.-of', and log/CFAR, a performance comparison of the cell averaging and 'greatest-of' CFAR techniques, and a summary of the results with recommendations for future effort. The performance of the two CFAR methods is independent of the detector. The cell averaging gives better detection performance in an interfering target environment, while the 'greatest-of' provides better false alarm regulation in a nonhomogeneous interference environment. (Author).