Search academic texts

Registration Number

0212U006515, 0111U009015 , R & D reports

Title

Development of methods for spatial filtering signals from magnetocardiograhic measurements

popup.stage_title

Head

Bytrim Oleksandr,

Registration Date

24-04-2012

Organization

Kharkov National University named after V.N. Karazin

popup.description2

Object of the research is magnetic fields from biological objects. The subject is algorithms for proc-essing data of measuring magnetic fields from biological objects by a system of sensors. The methods of studying is based on using optimal linear filtering (Viner's), building projectors in the multidimensional space of measured signal with help of singular decomposition and applying linear filters to input data in optimization methods that are based on minimizing quadratic discrepancy. Aim of the work is to develop methods for spatial filtering of the results of measuring magnetic field distribution, such that to separate the part of the measured signal that satisfy the following requirements: rapid decrease of sensitivity with distance; minimal loses of the useful signal from the local sources. The problem is considered regarding developing new methods for solving an inverse problem of mag-netostatics in magnetocardiographic measurements with concurrent filtering noise based on spatial filters, which allows making measurements in unshielded environment or by less expensive in manufacturing measurement systems based on magnetometers. The general problem statement is given for the magnetocardiographic meas-urements by system of sensors that consists of induction coils. The existing methods of filtering noise from ex-ternal sources are demonstrated that are based on coupling the coils into gradiometric structures. it is shown that using gradiometers can be considered as filtering by projecting onto a subspace in the signal space. The problem is stated about finding the optimal filters that would allow filtering signals from the heart at background of strong interference from external sources of magnetic field. Two possible schemes for building the spatial filter are considered: based on Taylor's expansion of the field in vicinity of the sensor system and building the pro-jectors onto the subspace of the corresponding Taylor harmonics, and building the subspaces that provide de-creasing sensitivity as a power law. The method is demonstrated for building such filters for a system of nonper-fect sensors that have, e.g. small spread in sensitivity or orientation, which is known due to calibration. The cal-culations have been carried out for revealing the potential of the proposed algorithm regarding reconstruction of parameters of a dipole source at the background interference from external sources. The high efficiency of such filtering is demonstrated. MAGNETOCARDIOGRAPHY, OPTIMAL FILTERING, INVERSE PROBLEMS, MAGNETOSTATICS

Product Description

popup.authors

Костев О.
Легенький М.
Шульга С.

popup.nrat_date

2020-04-02