Fault Finding in Antenna Array using Adaptive Neuro Fuzzy Inference System (ANFIS)

Abstract

Microstrip antenna has a planar structure and is easy to fabricate. However, it has a limitation of having low gain. This limitation can be overcome by using array of microstrip antennas. The applications of microstrip array antenna includes SONAR, RADAR, satellites, etc. Due to this high number of elements in an array, there is high probability of failure. This may affect the performance of array in terms of reduction in gain, change in radiation pattern, directivity etc. It is difficult to detect such elements for inaccessible situation like for space applications. Many researchers have used different techniques such as artificial neural network, bacteria foraging algorithm, genetic algorithms, etc. to diagnose the faulty element in an antenna array. The main aim of the current dissertation thesis is to find fault in antenna array using ANFIS [Adaptive Neuro-Fuzzy Inference System]. In order to diagnose the location of faulty element and to estimate the error in magnitude and phase of excitation current a 4x4 planar array antenna is designed using Ansoft Designer 4. To diagnose the location of faulty element, the radiation pattern of the faulty ele- ment is considered in the Ey plane. For =0 the radiation pattern remains same irrespective of the location of the faulty elements in the particular column. Similarly, for =90, the radiation pattern remains same irrespective of the location of the faulty elements in the particular row. Hence, =0 determines the column of the faulty element and =90 determines the row of the faulty element . For Proposed model, in case of diagnosis of magnitude error in excitation, magnitude and phase of the radiation pattern is taken as an input, while in case of phase error in excitation magnitude, phase and magnitude error are taken as an input. The results obtained through the proposed model are compared with the simulated data.