Multi-Adaptive Neuro-Fuzzy Inference System for Dielectric Properties of Oil Palm Fruitlets

Accurate dielectric models are required for proper sensing and characterization of materials especially for the purpose of quality control. In this work, a multi Adaptive Neuro-Fuzzy Inference System (ANFIS) was designed to model the complex permittivity of the mesocarps of oil palm fruitlets within the frequency range of 2-4GHz. The system consists of two ANFIS models with same sets of inputs; one ANFIS model for the dielectric constant and the other for the loss factor. Training data were obtained from laboratory microwave measurements with the aid of Vector Network Analyzer (VNA) and used for the ANFIS model. The evaluation of the performance of the model confirms the suitability of the multi ANFIS model for rapid and accurate determination of the dielectric properties of the fruitlets.