Determination Of The Dynamic Measurement Error Of Eit Systems

Numerous attempts have been made to improve image reconstruction algorithms in EIT. In contrast, the magnitude of measurement errors of an EIT system was not an exploratory focus in the past. Especially in absolute EIT (a-EIT) where the resistivity of the examined object is determined on an absolute scale the influence of these errors can not be ignored since systematic errors are not cancelled by dividing the raw data by reference data (cf. Sheffield back projection). We present very simple phantoms and provocative adapters for a 16 electrode system using adjacent drive pattern to determine the measurement errors. The trans-impedance of each phantom is determined by only one component and output voltage is constant for all combinations of current injection and measurement. Therefore the electrical properties of the phantoms can be determined fast and precisely. Input and output impedance can be chosen independently of output voltage similar to that of the human thorax. Parameters like the absolute error of each driving/sensing channel, signal to noise ratio in each channel and especially the several kinds of crosstalk can be determined. The EIT system runs at normal speed during the measurement which is important since dynamic properties of the system often influence the real errors. We exemplarily present the results of systematic errors of a Goe-MF II EIT system under two different simulated operational conditions. Although this system has been optimized for a sufficient signal to noise ratio we conclude that the absolute measurement errors will have to be reduced.