Research On Distributed CFAR Fusion Detection Algorithm

Detection algorithm of distributed CFAR is the fusion of CFAR technology and Muti-source information fusion technology. It has become key technologies of radar target detection today. Different forms of fusion method are used to improve the detection performance, expand the scope of observations. Around CFAR detection technology, applying network radars detection in a real environment as background, distributed CFAR detection algorithm based on hard threshold and fuzzy logic are studied in this paper. The main work includes the following respects:According to the mathematical model of clutter, zero memory nonlinearity is applied to produce Random sequence of four clutter distribution. Three corresponding CFAR algorithm are studied, such as:CA-CFAR、OS-CFAR technique based on Rayleigh distribution, MLH-CFAR technique based on weibull distribution, simulation and performance analysis for these algorithms are completed.For clutter in the actual environment is Non-Rayleigh, distributed CA-CFAR detection algorithm is designed in the background of weibull clutter. The probability of false alarm and detection of the fusion center is calculated in this paper. How factors of the shape parameter and the number of detectors affect the detection performance are analysis. As the improvement of this algorithm, Distributed LMSCA-CFAR detection algorithm is also designed, Then we compared the detection performance of the two algorithms.The application of Fuzzy logic in distributed CFAR algorithm is studied in this paper. We analysis the method for calculating and design principle of the membership function. We have solved the membership function of two detectors:PTM detector and MLH detector, Which solve the key problem for these detectors used in distributed detection system. Then we designed distributed OS-CFAR detection algorithm and distributed LMSCA-CFAR detection algorithm in weibull clutter. The detection performance of the algorithm is studied in Rayleigh distributed clutter and weibull distributed clutter. How factors of sliding window length and the number of detectors affect detection performance of the algorithms are simulations. We also compare the detection performance of the two algorithms.