A Method to Improve the Accuracy of the Bioelectrical Impedance Phase-Angle Measuring System

The bioelectric impedance phase difference (BIPA) is a characteristic quantity for the changes in the electrical properties of the tissue, obtained from the Bioelectrical Impedance Analysis (BIA). Nowadays, the need for an automatic, specialized and safe device to measure BIPA in medical facilities is very pressing. One method that has been proposed to measure this parameter is to use a multiplier and calculate the value of the phase difference between the two signals. The bias (mean) (+/- SD) of the measurement of the phase difference between two signals with programmed phases using this method was 0.5 +/- 0.2. Consequently, to improve the accuracy of the measurement results, this study was conducted to propose another signal processing method which is based on the comparator and the operating principle of the phase lock loop for the BIPA measurement system. This system still consists of an analog signal generator at 100kHz, a current source circuit. However, the signal processors which process signal received from the body consist of an analog-to-pulse converter and a phase-difference detector. The phase angle between the two signals is then calculated from the DC voltage obtained at the output of the phase-difference detector and compared with the programmed one. The phase difference between the two programmed signals ranges from 5 degrees to 8 degrees with an increment of 0.5 degrees. The obtained results show that the bias of the measurement performed by the system using the proposed method is -0.01, the precision was 0.07 and percentage error was 4.31%.

Automated summary

A new signal processing method based on the comparator and the operating principle of the phase lock loop was proposed for BIPA measurement, aiming to improve the accuracy of the results. By calculating the phase angle between two signals and comparing it with the programmed one, the system showed a bias of -0.01, precision of 0.07, and percentage error of 4.31%.