Multi-frequency EIT and TAS Hardware Development by luckboy


Multi-frequency EIT and TAS Hardware Development

More Info
									Multi-frequency EIT and TAS Hardware Development
T. I. Oh1, J. S. Lee1, E. J. Woo1, O. Kwon2 and J. K. Seo3
1 2 3

College of Electronics and Information, Kyung Hee University, Kyungki, Korea Department of Mathematics, Konkuk University, Seoul, Korea Department of Mathematics, Yonsei University, Seoul, Korea

ABSTRACT: We present the development of a multi-frequency EIT (electrical impedance tomography) and TAS (transadmittance scanner) system. Both systems share a common main platform based on a DSP and wireless serial RF connection to a PC. A digital waveform generator is used in both EIT and TAS producing sinusoidal waveforms with a chosen set of frequencies in the range of 0.1 to 500 kHz. In EIT, we implemented a balanced constant current source using the improved Howland circuit with stray capacitance compensation and a constant voltage source was used in TAS. The maximal number of voltmeters and therefore electrodes in EIT is 64 and multiplexers are used for current injections. TAS is equipped with 16 ammeters to accommodate a scan probe with multiplexers and a planar array of 320 disk electrodes for exit current sensing. In both systems, the demodulation uses the digital phase-sensitive demodulation technique with a nonuniform sampling method. Signal averaging and automatic gain control are also implemented in voltmeters and ammeters. We discuss the performance of the developed multi-frequency EIT and TAS system and their uses in future experimental studies. Keywords: EIT, TAS (trans-admittance scanner), multi-frequency

Bio-impedance or bio-admittance techniques such as biomedical electrical impedance tomography (EIT) have been studied as a diagnostic method to visualize functional images inside the human body [1]. When we inject current into an electrically conducting subject such as the human body through a pair of surface electrodes, the internal current pathway is determined by the conductivity, permittivity and geometry of the subject. Any local change of the conductivity or permittivity distribution results in a change of the internal current pathway whose effect is conveyed to boundary voltages. These boundary voltages are known to be insensitive to this local change and the relation between them is highly nonlinear. Therefore, the EIT image reconstruction problem is severely ill-posed and it requires a very accurate measurement of frequency-dependent impedance or admittance. For this reason, an impedance imaging system must measure boundary voltages with high accuracy in terms of amplitude, frequency and phase. In this paper, we describe a multi-frequency EIT and TAS system operating from 50Hz to 500kHz. In order to measure impedances in several frequencies, we implemented a multi-frequency waveform generator and a phase-sensitive demodulator inside FPGAs. We developed a multi-frequency EIT and TAS system. We summarize the system configuration of EIT and then describe the TAS system for detecting breast cancer. We also show some of our experimental results using saline phantoms.

2. METHODOLOGY 2.1 Design and implementation of multi-frequency EIT system

The developed EIT system shown in Fig. 1(a) consists of one current source, thirty two voltmeters, main controller with wireless RF data link to a PC and a control and image reconstruction software. This system has a radially symmetric architecture so that each channel has a similar characteristic. Fig. 1(b) is the developed 32 channel multi-frequency EIT system.

VM #3 VM #2 VM #1 USB RF RF DSP Controller Communic ation FPGA Clock & Phase0 Distribution VM #4 VM #... VM #... VM #... VM #... VM #62


VM #63 VM #64

Switch Network

Fig. 1. (a) Configuration of the EIT system. (b) Developed 32 channel multi-frequency EIT system.

We designed and implemented the multi-frequency constant current source including the FPGA-based waveform generator, voltage-to-current converter, and automatic calibration circuit. In order to generate sinusoidal data at various frequencies, we implemented an address generator which could change the time interval and order of output data from in FPGA. The FPGA receives two inputs (interval data, A and clock gap data, B) from the main controller. These data determine the frequency. The address generator outputs data at every (100 × A) /( B + 1) ns as shown in table 1. It can produce sinusoidal waveforms from 39.0625Hz to 500kHz. Fig. 2 shows a balanced current source using the improved Howland current pump circuits [2,3]. Multi-frequency digital voltmeters were implemented based on the digital phasesensitive demodulation technique [4].


cycle 1 2 … A

Address in ROM

0 -> A -> … -> 1000 - A



V/I Switching Network

1 -> A + 1-> … -> 1000 – A + 1 …

Control Logic


A - 1 -> 2 * A + 1

-> … -> 1000 - 1


Table 1. Multi-frequency data addressing

Fig. 2 Balanced current source in EIT system.

2.2 Design and implementation of multi-frequency TAS system
Multi-frequency TAS system uses the same main platform based on a DSP and wireless serial RF connection to a PC as in EIT. Fig. 3(a) shows the block diagram of 320channel TAS system. The developed 320-channel TAS consists of a handheld reference electrode, 320 channel scan probe with digital switches, multi-frequency constant voltage source using the same technique in EIT and 16 ammeters. Each ammeter contains a current-to-voltage converter and the same multifrequency phase-sensitive demodulator used in EIT system. Fig. 3(b) shows the developed TAS system and a cylindrical phantom with 300mm diameter and 500mm height. At the bottom of the phantom, we placed a disc-shaped reference

electrode with 40mm diameter.

Constant Voltage Source Digital Waveform Generator Voltage Source

Reference Electrode


Serialcom FPGA RF Data Link

Ammeter Ammeter Ammeter Demodulaor I/V Converter Sw itch

Scan Probe (Switches and Electrode)



Fig. 3 (a) Block diagram of 320-channel TAS system. (b) Developed 320-channel multi-frequency TAS system and cylindrical saline phantom.

3. RESULTS 3.1 EIT experiments using saline phantom
Fig. 4(a) shows an EIT saline phantom with a diameter of 200mm. It was filled with a saline with 0.04S/m conductivity. Inside the phantom, we placed an insulator made by acryl with 30mm diameter. Using the measured voltage data with and without the anomalies at 10, 50, 100 and 200kHz frequencies, we could reconstruct difference images shown in Fig. 4(b). Fig. 4(c) shows difference images between two frequencies. The reciprocity error of the measured data was about 4-6%.






Fig. 4 (a) Saline phantom with an insulator. (b) Difference images of the phantom in (a) before and after placing the objects at 10, 50, 100 and 200kHz. (c) Difference images of an acryl in the saline phantom between (c-1) 10 and 50kHz, (c-2) 10 and 100kHz, and (c-3) 10 and 200kHz.

3.2 TAS experiments using saline phantom
The TAS phantom in Fig. 3(b) was filled with a saline of 0.14S/m. Fig. 5(a) shows trans-admittance maps from the saline phantom with an insulator made by polypropylene and stainless steel at 10, 50, 100, and 200kHz frequency. Objects placed at 10mm depth from the top surface of the saline phantom where the scan probe was placed. Each object has 10mm side length. Imaginary image for the polypropylene object was changed significantly by varying the frequency from 5-200kHz as shown in Fig. 6. We found that the multi-frequency TAS does not require a reference data set using a homogeneous object. This will enhance the practical applicability of the method since such a reference data is not available.




(b) real


(c) real


(d) real


Fig. 5 Trans-admittance maps from the saline phantom with a cubic anomalies (insulator and conductor) at 10 mm depth at (a) 10, (b) 50, (c) 100 and (d) 200 kHz.

(a) real


(b) real

imaginary (c) real


Fig. 6 Trans-admittance maps of the phantom at two different frequencies. (a) 10 and 50kHz, (b) 100 and 50kHz , (c) 200 and 50kHz.

In this paper, we implemented the multi-frequency EIT and TAS system. We used FPGAs for waveform generations, demodulations, switch controls and data communications. The operating frequency of the systems ranges from 0.1 to 500 kHz. Using measured multi-frequency data at different frequencies, we could reconstruct difference images from saline phantoms with objects. We plan to improve the performance of the systems by using a better calibration and automatic gain control method. For the verification of the performance, we will conduct more experiments using phantoms with anomalies having frequency-dependent characteristics in their complex conductivity values.

Acknowledgement: This work was supported by the grant R11-2002-103 from the Korea Science and Engineering

[1] D. Holder, ed., Electrical Impedance Tomography: Methods, History and Applications, IOP Publishing, London, UK, 2005. [2] A. S. Ross, G. J. Saulnier, J. C. Newell, and D. Isaacson, “Current source design for electrical impedance tomography,” Physiol. Meas., vol. 24, pp. 509-516, 2003. [3] A. J. Wilson, P. Milnes, A. R. Waterworth, R. H. Smallwood, and B. H. Brown, “Mk3.5: a modular, multifrequency successor to the Mk3a EIS/EIT system,” Physiol. Meas., vol. 22, pp. 49-54, 2001. [4] R. D. Cook, G. J. Saulnier, D. G. Gisser, J. G. Goble, J. C. Newell, and D. Isaacson, “ACT3: a high-speed, highprecision electrical impedance tomography,” IEEE Trans. Biomed. Eng., vol. 41, pp. 713-722, 1994

To top