Real-time tuning of cavity filters by learning from human experience: A vector field approach

摘要

The technique of tuning a cavity filter is purely a rule of thumb: only experienced tuning engineer is competent to the task. However, with the great development of the communication industry and the rapid increasing of production capacity, the need for tuning technicians becomes urgent. It is meaningful to replace this traditional manual tuning task with some more advanced and automatic methods. We hereby propose a real-time computer-aided tuning method based on the vector field approximating approach, which can be applied in robotic tuning systems in the near future. In this paper, we first make a literature review on some previous intelligent cavity filter tuning solutions. Then the method of employing vector fields to represent the change of S-parameters is proposed. We provide concrete procedures to drive the S-parameters curves to approximate towards the target. In the end, we give the experimental results which validate the flexibility of the method.