A novel approach for parameter estimation of Fricke-Morse model using Differential Impedance Analysis
In this paper we presented a new approach for parameter estimation of Fricke-Morse model (2R-1C circuit) of biological cell. Proposed method is based on Differential Impedance Analysis and it was applied in parameter estimation of five electrical bioimped
- PDF / 232,454 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 56 Downloads / 155 Views
Univeersity of Banja Luka, L Faculty off Electrical Enginneering, Patre 5, 78000 Banja Luuka, Bosnia and Herzegovina Universiity of Novi Sad, Faculty of Techhnical Sciences, Trg T Dositeja Obradoviüa 6, 210000 Novi Sad, Reepublic of Serbiaa ar.simic, zdenka, z vlad do}@etfbl.n net, sgoran@u uns.ac.rs {mita
Ab bstract. In thiss paper we prresented a new w approach forr parrameter estimaation of Frickke-Morse modeel (2R-1C circuiit) of biologicaal cell. Proposeed method is baased on Differenttial Impedancee Analysis andd it was applieed in parameterr esttimation of five electrical bioimpedancess: Total Bodyy Coomposition, Truunk-Trunk, Arrm-Arm, Leg-L Leg and Respirattion Rate. Thee proposed meethod has beenn evaluated regarrding the influuence of the number n of meeasurement frequencies on the overall numerical accuracy and a processingg tim me. Obtained results r are coompared with Complex Nonn Linnear Least Squuares data fittting and it waas showed thaat preesented approaach is signifiicantly faster (ratio of processsing times deepends on the number n of meeasurement frequencies). Additiional advantagge of the propoosed method iss verry low compuutation complexxity (it is not iterative) so it i cann be easily impplemented in portable and auttonomous lowcosst microcontroller-based systtems for bioim mpedance measureement and paarameter estim mation of the Fricke-Morsee moodel in real-tim me.
In literrature, many stuudies can be foound in which parameter estimaation of Fricke-Morse model was used as diagnosis d tool. For example, exaamination of dielectric d propeerties of meat [3], in vivo humaan lung tissue bioimpedance analysis b [4], in viivo real-time myocardium tissue t characteerization [5], body y water estimaation [6], etc. Basically, if there is some chaange in tissuee from normall condition, vaalues of model paarameters will change from values obtaineed when tissue was in normal conndition.
Keeywords. Bioim mpedance, estimation, signal processing
Howev ver, some studdies suggest use u of more complex c models which w include fractional-ordeer circuits [7] as they shown much m better reepresentation of o experimentaally obtained biioimpedance data. d Complexx models havve been widely in nvestigated in many m fields off biology and biomedib cine [8, 9], 9 but they doo not provide explanations e reelated to the underrlying structurees and physicaal processes at the cellular leveel [10], thus Frricke-Morse model m is used by b many authors [1 11]. Moreover, parameter esttimation of succh models is usu ually based on complex noon-linear leastt-squares (CNLS) method, baseed on the Gaauss–Newton [12] or Levenberg rg–Marquardt [13] algorithm ms. The mainn disadvantages of such approaach are [6, 13, 14]: (a) CNLS S is time consumin ng process, (b)) CNLS requirres high qualitty initial guess for the model parrameters, (c) thhere is always a possibility of converging c intto a local miniimums, and (dd) CNLS non-conv vergence is a coommon issue. The diffferential impeedance analysis (DIA) is a teechnique which can
Data Loading...
