2
Artificial Neural Network in Microwave Cavity Filter Tuning
2.1 Introduction
The growth and progress in the mobile telecommunication industry drives a big demand for microwave filters. Looking from the perspective of rapidly developing industry one must face the technological obstacles limiting its growth. In the case of microwave filters the production process is very difficult because of the requirements imposed on the final product and the precision that is implied by these requirements. It turns out that even a modern assembly line cannot guarantee the satisfactory precision that is repeatable in the industrial scale. This is the basic reason for the situation where the microwave filter manufacturing process is now only partially automated. When the device leaves the automatic assembly line there is no guarantee that it satisfies the requirements assumed in the design process. It may be said that the filter is pre-tuned and waits for the intervention of the skilled human operator – this phase is called the tuning process. The filter's construction allows for some tuning elements (practically screws) that influence the resonant frequencies of cavities and values of couplings between them.
In the literature many different attitudes to the filter tuning process are described. The basic idea is to find the relations between ...
