Performance of the fuzzy VI-CFAR detector in non-homogeneous environments

In this paper, we applied the concept of fuzzy logic to improve the performance of the Variability Index Constant False Alarm Rate (VI-CFAR) detector. The VI-CFAR detector is an intelligent detector which uses data from the reference window to compute two second-order statistics (the variability index VI and the mean ratio MR) to select one of the three detectors, namely, the Cell Averaging-CFAR, the Greatest-Of-CFAR and the Smallest-Of-CFAR. The proposed fuzzy VI-CFAR consists of establishing a new detection strategy based on the fuzzification of the two statistics as well as the three detectors. Some inference rules are then used to select the appropriate fuzzy detector. The final decision is obtained through defuzzification using the centre of gravity formula. Simulation results show that the fuzzy VI-CFAR detector outperforms the classical VI-CFAR in homogeneous and non-homogeneous environments.