Штучний інтелект

Анотація: Intracranial pressure monitoring signal (ICP) obtained in Neuro Intensive Care Units, often contains a large amount of noise and outliers. These artifacts not only directly lead to false alarms in automatic alert systems of monitoring-diagnostic complex for controlled therapy patient, but they are also heavily pollute the main features of the signal, making it impossible to accurately predict the secondary damage to the brain caused by intracranial hypertension. This paper proposes an efficient on-line two-step purification method of physiological signals based on Hampel identifier and Kalman filtering. Initially, clinical measurement of ICP signal undergo pretreatment in which the identification signal data structure, and the estimated level of noise outliers are removed by the filter Hampel robust. Outlier points are replaced by the median of those points. Next, the correction of these points, noise removal and evaluation of the signal based on adaptive autoregressive (AAR) model using the Kalman filter and its associated RTS (Rauch-Tung-Striebel) smoothing filter are performed. Adjustable parameters of the proposed filtering method is the half-width moving window, the threshold for detecting outliers and order AAR model.

Ключові слова: Hampel filter, Kalman filtering, adaptive AR model, outliers, intracranial pressure

Посилання:

1. Hampel F.R. The Breakdown Points of the Mean Combined With Some Rejection Rules / F.R Hampel // Technometrics. - 1985. - V. 27. - P. 95 - 107.
2. Martin, R. D. Robust-resistant spectrum estimation / R. D Martin, D. J Thomson // Proceeding of the IEEE. - 1982 - V. 70. - P. 1097-1115.
3. H. Liu, S. Shah, W. Jiang. On-line outlier detection and data cleaning. // Computers and Chemical Engineering. - 2004. - V. 28.- P. 1635-1647.
4. Data cleaning for dynamic modeling and control / R. K. Pearson // European Control Conference, ETH. - 1999. - Zurich, Switzerland.
5. F.R. Hampel, .E.M. Ronchetti, P.J. Rousseeuw, W.A. Stahel. Robuts Statistics. The Approach Based on Influence Functions. John Wiley & Sjns, Inc., New York 1986. 502 p.
6. Pearson, R. K. (). Outliers in process modeling and identification / R. K Pearson // IEEE Transactions On Control Systems Technology. - 2002. - V. 10. - P. 55–63.
7. Vaseghi Saeed V. Advanced digital signal processing and noise reduction / Saeed Vaseghi. — 4th ed., JohnWiley & Sons Ltd, United Kingdom. - 2008. - 514 p.
8. Schlögl A. Adaptive Methods in BCI Research - An Introductory Tutorial / A. Schlögl, C. Vidaurre, K.- R. Müller // In "Brain Computer Interfaces - Revolutionizing Human-Computer Interfaces" Springer. - 2010. - P. 331-355.
9. Drobotko D.V. Monitoring System and Morphology of Non-stationary Signals of Intracranial Pressure / D.V. Drobotko, I.V. Kachur, V.F. Drobotko, G.A. Gorodnik // Artificial intelligence. – 2010. - № 4. – С. 651-661.
10. . Online ICP Forecast for Patients with Traumatic Brain Injury / F. Zhang, M. Feng, L.Y. Loy, Z. Zhang, C. Guan // In Proc. ICPR-2012. - 2012. P. 37-40. -Tsukuba, Japan.
11. Artifact removal from intracranial pressure monitoring signals: a robust solution with signal decomposition / M. Feng, L. Y. Loy, F. Zhang, and C. Guan // in Proc. IEEE EMBC'11. - 2011. - P. 797 - 801. - Boston, USA.
12. Artifact Correction with Robust Statistics for Non-Stationary Intracranial Pressure Signal Monitoring / M. Feng, L. Y. Loy, K. Sim, C. Phua, F. Zhang, C. Guan // In Proc. ICPR-2012. - 2012. - P. 557-560. - Tsukuba, Japan
13. Neurophysiological monitoring of intensive therapy in the sharp period of a heavy craniocereberal trauma. Medicine of urgent conditions / V.I. Cherny, G.A. Gorodnik, A.M. Kardash [etc.] // Specialised scientificallypractical magazine. – 2008. – № 2 (15). – P. 72-76. .
14. ICP Monitoring Codman. http://freepdfdb.com/pdf/icp-monitoring-codman.
15. Allen D. P.. Suppression of deep brain stimulation artifacts from the electroencephalogram by frequencydomain Hampel filtering / D. P. Allen, E. L. Stegemöller, C. Zadikoff, J. M. Rosenow, C. D. MacKinnon // Clin Neurophysiol. - 2010. - V.121, № 8. - Р. 1227–1232.
16. Schlögl A. The Electroencephalogram and the Adaptive Autoregressive Model: Theory and Applications. Dissertation, (Dr. techn.) / Schlögl A. – April 2000, Graz. – 83 p.