Signal processing and event detection of hip implant acoustic emissions

详细信息    查看全文

• 作者：Geoffrey W. Rodgers^a ; ^{geoff.rodgers@canterbury.ac.nz} ; Roger Welsh^a ; Laura J. King^a ; Anthony J. FitzPatrick^a ; Tim. B.F. Woodfield^b ; Gary J. Hooper^b
• 关键词：Biomedical systems ; Vibration measurement ; Vibration monitoring ; Ultrasonic transducers ; Signal analysis ; Frequency spectrum ; Acoustic emissions ; Orthopaedic implants ; Wavelet transform
• 刊名：Control Engineering Practice
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：58
• 期：Complete
• 页码：287-297
• 全文大小：1644 K
• 卷排序：58

文摘

Acoustic Emission (AE) monitoring is a non-invasive method for determining the condition of Total Joint Replacement (TJR) implants. In this study, recordings from both in-vivo patient tests and in-vitro implant measurements are compared, with key acoustic frequency profiles identified. Event detection, removal of ambient noise, and characterisation of audible squeaking is investigated. Fundamental implant frequencies from squeaking are within 2–5 kHz, with several harmonics also present.Several event detection methods are investigated, including a statistical method based upon sound intensity, wavelet analysis and a Root-Mean-Squared (RMS) variation method. All methods can effectively identify most events, with the wavelet approach being the most effective, and the RMS method the most computationally demanding. The identified events are then utilised to determine the baseline ambient noise characteristics, which are then removed from the signal to increase the clarity of frequency peaks.Comparisons are made between the in-vivo tests and the subsequent in-vitro results, with a clear correlation between the two test methods. In-vivo signals were notably lower magnitude than recordings from subsequent in-vitro laboratory tests made on the implant directly. However, the peak frequency range was closely similar in both cases, indicating that signal attenuation through human soft tissue reduces signal amplitude. The test results and analysis methods presented are key in the development of future data collection and analysis for the development of a full AE-based clinical diagnostic tool.