Automatic identification of otological drilling faults: an intelligent recognition algorithm

Abstract

BACKGROUND: This article presents an intelligent recognition algorithm that can recognize milling states of the otological drill by fusing multi-sensor information. METHODS: An otological drill was modified by the addition of sensors. The algorithm was designed according to features of the milling process and is composed of a characteristic curve, an adaptive filter and a rule base. The characteristic curve can weaken the impact of the unstable normal milling process and reserve the features of drilling faults. The adaptive filter is capable of suppressing interference in the characteristic curve by fusing multi-sensor information. The rule base can identify drilling faults through the filtering result data. RESULTS: The experiments were repeated on fresh porcine scapulas, including normal milling and two drilling faults. The algorithm has high rates of identification. CONCLUSIONS: This study shows that the intelligent recognition algorithm can identify drilling faults under interference conditions.