AUTOMATED INSTRUMENT FOR HUMAN EAR SURGICAL PROCEDURE BASED ON MACHINE VISION AND SENSING FEEDBACK

Abstract

In this thesis, two versions of the surgical instrument are presented to deal with the human ear diseases happening in the middle and inner ear respectively. The first version leveraging on Kalman-filter based visual servo, fuzzy logical gain scheduled PI control and multi-sensing switching mechanism is able to operate the surgery of inserting a grommet tube on the patient’s tympanic membrane from any initial position outside the ear automatically. Based on the first version, the second version which has the access to the middle ear cavity can orchestrate a sequence of fine movements of the working channel to reach the targeted spot for drug delivery to the inner ear. Making using of a synchronized set of algorithms including image feature extraction, boosting classification and 3D image reconstruction, the desirable spot for drug delivery can be accurately positioned. Pre-clinical tests have been carried out to validate the proposed instruments.