An Efficient Assistance in Cavity Filter Tuning Using Filter Screw Classification

The research community is primarily focused on automating the filter tuning process. Effectively distinguishing the tuning screws from the mounting screws present on the surface of cavity filters is one of the key factors to improving the efficiency of automated filter tuning process. Since the tuning state of the filter is altered only by rotating the tuning screws, it is imperative to detect and localize the position of these tuning screws. This paper presents a supervised Machine Learning (ML)-based approach to differentiate between the tuning screws and mounting screws of a cavity filter. The proposed methodology underwent evaluation on a commercial cavity filter, achieving an impressive precision of 96.66%. Other state-of-the-art supervised learning-based algorithms have been tested against the Support Vector Machine (SVM) classifier. The empirical findings unequivocally demonstrate the superiority of the proposed methodology in comparison to alternative approaches. The findings of this research work can assist in automating the filter tuning process.