| Although several modeling techniques have been used to model indoor radio wave propagation and coupling patterns, to date no efficient model exists that calculates indoor-outdoor radio wave propagations on commercial aircraft. Due to the complexity of an aircraft structure, with the additive introduction of creeping wave phenomenon and unknown back-door propagation values from the exterior aircraft antenna to the avionics bay, numerical modeling approaches using Method of Moments (MoM) or Finite Difference Time Domain (FDTD) prove too complex with limitations. This dissertation presents an expert neuro-fuzzy (NF) model for Interference pathloss (IPL) predictions inside an Airbus 320 (A320) airplane, for radio systems from 75 to 1585 MHz. This novel model generates IPL pattern through fuzzy logic, incorporating linear expert knowledge into the patterns. The model also uses feed-forward neural networks to derive meanings from complicated or imprecise data, extract patterns and detect trends in the IPL data that are too complex to be noticed by either humans or other computer techniques. Unlike previous approaches, the model presented is robust in incorporating both low to high band frequencies. It is also computationally efficient and reliable. |
