J. Cent. South Univ. (2016) 23: 1198-1207

DOI: 10.1007/s11771-016-0369-4

RF characteristics of wireless capsule endoscopy in human body

WANG Jing-chen(王璟琛)¹, WANG Zhao(王炤)¹, Mark Leach¹, LEE Sanghyuk¹, Eng Gee Lim¹, HUANG Yi²

1. Department of Electronic and Electrical Engineering, Xi’an Jiaotong-Liverpool University, Suzhou 215123, China;

2. Department of Electrical Engineering and Electronics, The University of Liverpool, Liverpool, L69 3GJ, UK

Central South University Press and Springer-Verlag Berlin Heidelberg 2016

Abstract: The wireless capsule endoscope, as a small electronic device, has conquered some limitations of traditional wired diagnosing tools, such as the uncomfortableness of the cables for the patient and the inability to examine the very convoluted small intestine section. However, this technique is still encountering a lot of practical challenges and is looking for feasible improvements. This work investigates the RF performance of the wireless capsule endoscope system by studying the electromagnetic (EM) wave propagation within the human body. A wireless capsule endoscopy transmission channel model is constructed to serve the purpose of investigating signal attenuations according to the relative position between the transmitter and the receiver. Within 300-500MHz, the S₂₁ results are regular and do not display any sudden changes, which allows a suitable expression to be derived for S₂₁ in terms of frequency and offset. The results provide useful information for capsule localization.

Key words: wireless capsule endoscopy (WCE); electromagnetic wave propagation; signal attenuations; positioning; transmission channel

1 Introduction

The history of using endoscopes to examine human internal organs can be referenced back to the 19th century [1] when a German scientist first developed a ‘Lichtleiter’ to examine the human bladder and bowel utilizing candle light. Since then, many different types of endoscope have been developed to help doctors scan a variety of human internal organs. Timely detection and diagnoses of medical problems are extremely important, as administering treatment at the earliest possible stage generally leads to higher success rates in curing patients [2]. Of particular interest in this work is the gastro- intestinal (GI) tract, for which the majority of related cancers are curable, when discovered at an early-stage. Traditional surgical endoscopes can be divided into two main branches, gastroscopy and colonoscopy. The first is used to check the stomach and the latter the colon. Over the last two decades, various developments in technology have allowed the combination of these two types of endoscope to be explored, leading to the birth of the capsule endoscope. Capsule endoscopy as shown in Fig.1 is non-invasive and offers patients improved comfort over traditional endoscopes. They can be used to examine more deeply inside the gastrointestinal (GI) tract of the human body, than can currently be achieved using the existing wired endoscopes.

Wireless capsule endoscopy (WCE), utilises a capsule-shaped small electronic device containing a video camera, LED lights, a power source and a wireless transmitter. It can be used to detect digestive system diseases (in, for example, the duodenum, jejunum or ileum). There are many different types of wireless capsule endoscopes, most of which have been developed and commercially manufactured [3]. Furthermore, some capsule endoscopies with state-of-the-art locomotion systems are summarized in Ref. [4]. However, there are still several drawbacks limiting the application of WCE. Firstly, the physiological data collected, such as GI tract images, are insufficient for clinical diagnosis without accompanying capsule position information. Secondly, most capsules are powered by an internal battery cell that in turn restricts capsule miniaturization. Finally, the random orientation of the capsule as it travels through the GI tract means that a continuous communication link to the outside of the body is not maintained [5-6].

This work investigates the RF performance of the WCE communication system, studying the electromagnetic (EM) wave propagation characteristics within the WCE system based on the empirical pass loss model, and evaluating signal attenuation due to the distance between the transmitter and receiver [7-8]. This investigation serves to examine signal attenuation through the human body as well as determining a method for obtaining capsule position. There are two main reasons for this proposed research. Firstly, EM energy is absorbed by organs as it propagates through the human body, which can lead to large signal distortion. The human body is a frequency dispersive system, with frequency dependent parameters (permittivity and conductivity) [9] that influence the electric and magnetic performance of the signal as it traverses the channel. Since the properties are frequency dependent, then wide- band signals present a significant design challenge. Simulations of signal propagation in such a system require a model of the human body, whose physical structure is extremely complex and even more so in the view of its dielectric properties.

Fig. 1 Human GI tract and wireless endoscopy capsule [3]

Secondly, it is proposed that the position of the capsule within the human body can be obtained by studying the EM wave propagation characteristics of the system, without the use of additional sensors. As well as yielding the capsule location and orientation information, it will save space within the capsule for other components.

In this work, simulations of wave propagation through the human body are developed and analysed, to examine signal distortion due to the transmission channel for the WCE system. This also serves to characterise the communication process over varying distances, to pave the way for positioning information to be obtained. Multiple excitation signals are applied to test the system including a normal Gaussian signal, a 10th order Gaussian wavelet and a binary amplitude shift keying (BASK) signal. The modelling process is described and results are shown.

2 WCE communication model

In order to study wave propagation in a WCE system, a model of the system must first be developed, and in this case the model consists of a transmitter, a receiver and the communication channel. To better understand the system as a whole, the communication channel, the transmitter and the receiver will be discussed separately.

2.1 EM wave propagation channel environment

The anatomy of the human abdomen is structurally highly complex from a dielectric perspective. When a capsule travels inside the small intestine, the body materials surrounding it consist of muscle, fat and skin, in a variety of shapes and each with varying dielectric properties. The material closest to the capsule presents the greatest influence from an electromagnetic viewpoint. The highly complex model that would be required to describe the inhomogeneous structure of the human body can therefore be simplified to a homogeneous body model based on muscle material (that lying closest to the capsule). The accepted dielectric properties of muscle material, obtained from literature, suggests a permittivity of e_r=56 and a conductivity of s=0.83 S/m [10] at 400 MHz, which is the operating frequency of the transmitter-receiver system. The shape of the body model can be simplified to a cylinder or ellipsoid, as shown in Fig.2, the chosen body model consists of a cylinder with an oval cross section. The shorter radius of 100 mm lies in the y axis and longer radius of 150mm in the x axis. The size has been chosen based on the needs of the study; extra layers of different tissues could easily be added in future investigations.

Fig. 2 Simplified human body model (homogeneous with ε_r=56 and σ=0.83 S/m)

2.2 Transmitting and receiving antennas

To implement the communication system, a suitable transmitting antenna (T_x) and receiving antenna (R_x) are selected to operate through the human body environment. The WCE antenna should be designed to be insensitive to human tissue influences, since it transmits from inside the body. The electrical properties of the human body mean that it is classified as a lossy dielectric material and hence it absorbs power, decreasing a waves’ energy as it propagates. A wide bandwidth signal is required to transmit high resolution images, which contain large amounts of data; hence the transmitting antenna needs to be wideband. Detection of the transmitted signal is preferred to be independent of transmitter position or orientation and hence, the transmitting antenna should have an omni-directional radiation pattern. On the basis of Friis’s formula [11], the total loss between transmitter and receiver is calculated in relation to the distance between the transmitting and the receiving antennas, which in this case is determined to be 150mm (the largest dimension of the body model in Fig.2). Using the dielectric properties of the muscle tissue, calculation shows that the minimum total loss is achieved when the operating frequency is in the range of 400-600MHz [12-13].

Extensive investigations have been performed on the development of WCE system antennas. One type of antenna that has been studied and applied is the confocal antenna, which is printed on, or attached to, the inner or outer surface of the capsule shell [14-18]. These antennas do not occupy space within the capsule, saving precious spatial capacity for other devices or larger batteries. However, these antennas are exposed to the surrounding human body tissues, which can easily lead to antenna frequency detuning. Another important category of antenna is the spiral antenna, including the basic single arm circular spiral antenna [19] and its modified versions, such as the dual circular spiral [20], conical spiral [21] and fat arm spiral [22]. Spiral antennas offer relatively wide bandwidths, which provide a margin for any frequency shifting caused by the varying properties of the human body tissues as the capsule moves.

In this work, a spiral antenna developed by KWAK et al [19] is selected as the transmitter, as shown in Fig.3. This antenna module is constructed from coaxial cable, a ground plane, a substrate layer, a metal plane (spiral) and an air gap. Seen from the top, the antenna begins at a central circular patch to which the coaxial cable is attached. The spiral starts at a predetermined radius and is connected to the centre by a straight metal section. The asymmetric structure makes the radiation pattern in this plane worthy of some attention.

In previous works, various receiving antennas have also been investigated. Simulation results show that the PIFA antenna appears to be too large to form a suitable far-field, and the tested UWB type antenna’s phase responses are extremely different from those of the spiral transmitting antenna, which would result in large signal delay. For the purpose of matching, the sizes and phases of S₂₁ for the T_x/R_x pairs, the application of a similar spiral antenna at the receiver is suggested (as shown in Fig.4).

Fig. 3 Top (a) and side (b) view of spiral antenna [19]

Fig. 4 Transmitter (T_x) and receiver (R_x) antenna pair

From previous studies, the resonances of the T_x and R_x antennas can undergo shifts to higher or lower frequency regions when the distance between the antennas changes [23]. After optimization, an overlapping of the operational bandwidths of the T_x and R_x antennas of 85MHz is achieved. The operating frequency of the system is selected as 403MHz.

2.3 Two-port network for WCE communication system

The communication system constituted by the T_x, intermediate material and R_x can be considered as a two- port network. Therefore, scatting parameters can be used to analyse the system, as shown in Fig.5. S₁₁ is the return loss, used to determine the channel bandwidth. S₂₁ is the ratio of voltage transmitted to Port2 to the voltage sent from Port1, known as the forward voltage gain.

Fig. 5 WCE communication system (a) and scattering parameters (b)

In the simulation, the T_x and R_x antennas are positioned such that the T_x is inside the body model and the R_x outside separated by a distance of r=100mm. The terminals of the T_x and R_x antenna cables are set as Port1 and Port2, respectively. The scattering parameters of this communication system are shown in Fig.6. The return losses S₁₁ and S₂₂ show approximately 120MHz of overlapping operational bandwidth from 341MHz to 462MHz and the operating frequency of both is close to the design frequency of 403MHz.

Fig. 6 Scattering parameters of T_x and R_x pair in human body communication system

3 EM wave propagation evaluations

3.1 Relative angle position between T_x and R_x

The spiral antenna is not a symmetrical structure, which means that the antenna radiation pattern (RP) and hence the angular distribution of power will not be symmetrical. To investigate the effect of this system, a model is used to find signal attenuation with respect to the T_x/R_x angle. The body model is reshaped to a circle with a radius of r=100mm and the receiver is rotated about the cylinder in 45°increments (illustrated in Fig.7), where position A represents 0° with the antennas boresights aligned. The simulation results for S₂₁ at 403MHz are collected and compared with the transmitter’s radiation pattern over the examined surface.

Fig. 7 Testing points for evaluating RP effects

The calculated forward transmission parameters (S₂₁) at 403MHz are shown in Fig.8. They are compared with the transmitter’s directivity radiation pattern calculated at different angles in the examined plane. From Fig.8, the S₂₁ performance of the system as a function of angle between T_x and R_x antenna is almost proportional to that of the T_x antenna radiation pattern. For this system, the forward voltage gain is around -80dB with a ±4dB variation.

The angle-dependent forward voltage gain is an important factor influencing the calculation of capsule location and therefore needs to be taken into account while performing localization estimation.

3.2 Transfer functions and S₂₁

The simulated S₂₁ results need to be examined as to whether they could be considered as a transfer function. The discrete Fourier transform (DFT) is applied to both the input and output port signals enabling the transfer function to be found. Comparison of this transfer function with the S₂₁ results obtained from CST will allow determination of the validity of using the S₂₁ results. Limited by the simulated time domain resolution of the signal, the number of points within 1GHz is small, but the general trend of the transfer function could be obtained.

Fig. 8 Radiation pattern’s 2D plot (a) and S₂₁ (b) vs angle at 403 MHz

Figure9 versus compares the transfer function and S₂₁ versus frequency. It can be seen that the two plots follow the same general trend over the frequency range of 300-500MHz in both absolute value and phase. Therefore, a transfer function as a function of the offset distance and frequency could be achieved by fitting the simulated S₂₁ results using various offsets at the simulated frequency points.

3.3 Relative distance between T_x and R_x

Having established a link between S₂₁ and the system transfer function, the effect of different offsets between transmitter and receiver in terms of their effect on S₂₁ is now investigated. In keeping with the system parameters in terms of capsule localization, signal transmission distances are swept from r′=0-160mm in 20mm steps. The position of the R_x antenna is fixed and the T_x antenna is moved over the defined range of distances through the human body phantom in the plane of the R_x antenna. Figure10 shows how the data are obtained.

Fig. 9 Gaussian’s S₂₁ vs transfer function

At each offset the model was used to obtain S₂₁ over a range of frequencies; the results are listed in Table 1.

Fig. 10 Layouts of T_x and R_x:

Table 1 S₂₁ at selected frequencies with various offsets from T_x to R_x

From Table 1, it can be seen that over the frequency range of interest (300-500MHz), the values for S₂₁ are regular and do not show any abrupt irregularities. These groups of data can therefore be used to fit the channel transfer function and will be used in the evaluation of the communication system in the next section.

4 Evaluation of communication system

4.1 Transfer function prediction with various offsets

Fitting the transfer function is meaningful. If the transmitted signal’s amplitude is known, when the received signal’s amplitude is detected, the level of attenuation can be calculated. For example, if the attenuation is found to be -80dB when operating at 403MHz, then by referring to Fig.11, the offset, r′, can be determined as 100mm, and this method can be used to determine capsule position.

Table 1 shows S₂₁ for different offsets and frequency ranges. The task is to find a suitable expression for S₂₁ in terms of frequency and offset. The task of curve fitting is accomplished by first starting with the theoretical relationship:

                            (1)

          (2)

where so V₁ is the source potential and

         (3)

The potential at one point (not the source) equals the integral of the electric field, and is inversely proportional to the distance from the reference point to the source. If V₂ is the potential at the receiving end, and V₁ is the signal transmitted from the source, then S₂₁ can be expressed as Eq. (3). However, if we fit the curve of S₂₁ with a function related to r^-1 (the offset distance), then it is very difficult to find suitable parameters that provide a confidence bound larger than 95%. Other methods of curve fitting the simulated S₂₁ dataset have been attempted and a possible transfer function is achieved by applying the following correction factors:

                            (4)

where r is the offset, mm, and f is frequency, Hz.

Figure 11 reveals that in comparing the S₂₁ results to the transfer function (with correction factors from Eq. (4)) the results are highly accurate. It is of note that these results are obtained using a Gaussian excitation signal.

Fig. 11 Fitted curve

4.2 System performance with BASK signal

Amplitude shift keying (ASK) is one of the most favoured digital signalling modulation techniques and has been applied in most WCE systems. Compared with other methods like frequency shift keying (FSK), it is highly power efficient, which makes it suitable for power limited systems such as WCE. Therefore, this system is tested with a practical binary ASK (BASK) port signal.

The data size of the pictures taken by the image sensor of the capsule is 255×255 pixels and every pixel occupies 24 bits; also the frame rate is 2Hz and RGB colour images have 3 dimensions, and the minimum bit rate is given by

          (5)

And based on the Shannon’s channel capacity theory, there is

    (6)

The worst case scenario in terms of bandwidth is 85MHz. The temperature inside the human body ranges from 36.9° to 37.9°, with a mean value of nearly 36.5°, and

N₀B=kT×BW=274.5×1.38×10^-23×85×10⁶=3.22×10^-13 (7)

To ensure that the channel can transmit data without error, and also make a trade-off between bandwidth efficiency and energy efficiency, assume that R_b is 360 times smaller than the channel capacity C.

Using BASK modulation, the signal is given by m(t)×cos(ω_ct), where

         (8)

                             (9)

Since for an ASK signal the bit energy is typically E_b=2.952×10^-8, then typical BASK signals are given by

                 (10)

BASK signals consisting of two bits have been analysed using CST; the logic ‘1’ and ‘0’ potions of Figs.12 and 13 show the input and output signals obtained from the simulations.

Fig. 12 Input BASK signal at Port 1

Fig. 13 Output signal received at Port 2

A sequence of ‘0’s and ‘1’s modulated as BASK are shown in Fig.14. The corresponding received signal following transmission through the model is shown in Fig.15.

The frequency response of the whole system is referred to as the transfer function, which can be calculated from the transmitted and received signal. It is calculated and compared with the simulated S₂₁ in Fig.16.

4.3 System performance using a wavelet Gaussian signal

The bit duration of a BASK signal (1/R_b) is much longer than the period of a standard Gaussian pulse, which leads to very slow simulations in CST. As a compromise, a 10th order Gaussian wavelet signal is used.

The Gaussian wavelet signal is a derivative of the standard Gaussian pulse and can be generated in MATLAB using the function: ‘gauswavf’, as shown in Fig.17. Wavelet signals have high accuracy in both high and low frequency ranges, and therefore, are also referred to as the ‘mathematics microscope’ as they can meet the requirements of multiple time and frequency domain analyses. The 10th order Gaussian wavelet signals are chosen to be used in this simulation.

Fig. 14 Message signal (a) and modulated BASK signal (b)

Fig. 15 Received signal in time domain

The transmitted and received signals, based on the 10th order Gaussian signal, are shown in Figs.18 and 19, respectively.

The simulated S₂₁ with Gaussian applied wavelet signal, illustrated in Fig.20, fits the transfer function for almost all frequency ranges investigated, with the exception of the frequency interval below 100MHz. The results are better than that of the transfer functionachieved when a standard Gaussian excitation signal is applied.

Fig. 16 BASK’s S₂₁ vs transfer function

Fig. 17 Shape of 10th order Gaussian wavelet signal in time domain

Fig. 18 Input 10th order Gaussian wavelet signal at Port 1

Fig. 19 Output signal received at Port 2

Fig. 20 10th order Gaussian wavelet’s S₂₁ vs transfer function

Increasing the simulation’s sampling period could increase the accuracy in that range. However, in spite of the lack of accuracy in the low frequency range, the Gaussian wavelet’s simulation results are very close to the results of BASK in the frequency spectrum above 200MHz for an offset of 100mm. S₂₁ for other offsets needs to be investigated in the future for verification. Providing that the wavelet’s results are close to those for the BASK signal for all ranges of offset, the Gaussian wavelet could be used to replace BASK signals, saving simulation time. It is important to note that there is no implication that the 10th order Gaussian wavelets are better than Gaussian port signals or BASK signals, and they merely lead to faster simulations.

4.4 Comparison of excitation signals

Figure 21 shows CST S₂₁ simulation results for three different excitation signals investigated. It can be seen that the three curves follow the same trend for frequencies larger than100MHz. However, the standard Gaussian result is less well fitted to the BASK signal than the 10th order Gaussian wavelet.

Comparison of the simulation features between the standard Gaussian signal, the 10th order Gaussian wavelet signal and BASK signals is given in Table 2. In terms of simulation time, it could be found that BASK simulations have the highest cost, almost 100 times longer than the normal Gaussian and 10th order Gaussian wavelet simulations. In terms of frequency range, the bandwidth of the standard Gaussian, 10th order Gaussian wavelets and BASK have been found to be approximately 150MHz, 19MHz (2 times the data rate) and less than 1Hz, respectively.

Fig. 21 S₂₁ of Gaussian, Gaussian wavelet and BASK signals

Table 2 Comparison of simulation features among normal Gaussian signal, 10th order Gaussian wavelet signal and BASK signals

As shown in the previous sections, the accuracy of the BASK and Gaussian wavelet signal’s S₂₁ simulation results are better than those of the normal Gaussian signal. Thus, S₂₁ achieved using these two types of excitation signals can describe the transfer function better than the S₂₁ achieved using the standard Gaussian signal as the input.

By making trade-offs between simulation time, frequency range and accuracy of results, the suitable input port signal type can be chosen based on the simulation requirements.

5 Conclusions

1) With the propagation information, it has been shown that the human body is a very lossy material causing huge signal attenuation. The signal attenuation rate is related to both the distance between the two antennas and the direction of the antennas.

2) To reduce bit error rate, the modulation method and the data rate should be chosen carefully.

3) The investigation results could provide information regarding capsule localization at various signal transmission distances.

4) Further research needs to be carried out on various offset distances with respect to the presented results.

Acknowledgment

The authors would like to express their sincere gratitude to CST AG for providing the CST STUDIO SUITE electromagnetic simulation software package under the China Key University Promotion Program, and the comprehensive support on it.

References

[1] Endoscopy [M/OL]. From Wikipedia, the free encyclopedia: http://en.wikipedia.org/wiki/Endoscopy. 2011.

[2] THOTAHEWA K M S, REDOUTE J M, M, YUCE R. A UWB wireless capsule endoscopy device [C]// Int IEEE Eng Med Biol Soc. Chicago, USA, 2014: 6977–6980.

[3] LIM E G, WANG J C, WANG Z, TILLO T, MAN K L, ZHANG N. Wireless capsule endoscopy [C]// International Multi Conference of Engineers and Computer Scientists. Hong Kong, 2013: 700-701.

[4] LIU L, TOWFIGHIAN S, HILA A. A review of locomotion systems for capsule endoscopy [J]. Biomedical Engineering, 2015, 8: 138-151.

[5] FIREMAN Z, KOPELMAN Y. New frontiers in capsule endoscopy [J]. Gastroenterology Hepatology, 2007, 22(8): 1174-1177.

[6] WANG A, BANERJEE S, BARTH B A, BHAT Y M, CHAUHAN S, GOTTLIEB K T, KONDA V, MAPLE J T, MURAD F, PFAU P R, PLESKOW D K, SIDDIQUI U D, TOKAR J L, RODRIGUEZ S A. Wireless capsule endoscopy [J]. Gastrointestinal Endoscopy, 2013, 78(6): 805-815.

[7] WANG Z, LIM E G, ZHANG M, WANG J C, TILLO T, CHEN J H. Electromagnetic wave propagation of wireless capsule antennas in the human body [C]// Antennas & Propagation (ISAP). Nanjing, 2013: 625-628.

[8] THOTAHEWA K M S, REDOUTE J M, YUCE M R. Propagation, power absorption, and temperature analysis of UWB wireless capsule endoscopy devices operating in the human body [J]. Microwave Theory and Techniques, 2015, 63(11): 3823-3833.

[9] The Finite Integration Technique [M/OL]. CST Computer Simulation Technology AG. http://www.cst.com/Content/Products/MWS/FIT. aspx.

[10] LIM E G, WANG Z, CHEN J H, TILLO T, MAN K L. Investigation of EM wave propagation of the wireless capsule in human body [C]// Design & Test Symposium. Rostov-on-Don, Russia, 2013: 1-4.

[11] KWAK S I, CHANG K, YOON Y J. The helical antenna for the capsule endoscope system [C]// Proc IEEE Antennas Propag Symp. Washington DC, USA, 2005: 804-807.

[12] ABDILL L, SAMMUT C, MANGION L Z. Dielectric properties of muscle and liver from 500 MHz-40 GHz [J]. Electromagn Biol Med, 2013, 32(2): 244-252.

[13] LIM E G, WANG J C, WANG Z, TILLO T, MAN K L. The UHF Band in-body antennas for wireless capsule endoscopy [J]. Engineering Letters, 2013, 21(2): 72-80.

[14] KIM K, YUN S, LEE S, NAM S, YOON Y J, CHEON C. A design of a high-speed and high-efficiency capsule endoscopy system [J]. IEEE Transactions On Biomedical Engineering, 2012, 59(4): 1005- 1011.

[15] WANG J C, LIM E G, WANG Z, HUANG Y, TILLO T, ZHANG M, ALRAWASHDEH R. UWB planar antennas for wireless capsule endoscopy [C]// International Workshop on Antenna Technology. Karlsruhe, Germany, 2013: 340-343.

[16] ALRAWASHDEH R, HUANG Y, CAO Ping, LIM E G. A new small conformal antenna for capsule endoscopy [C]// 7th European Conference on Antenna and Propagation. Gothenburg, 2013: 220-223.

[17] LIM E G, WANG Z, XIA T Q, LEACH M, MAN K L. The new wrappable wireless capsule antennas [C]// ISOCC. Jeju, Korea, 2014: 177-178.

[18] LIM E G, WANG J C, WANG Z, JUANS G, TILLO T, MAN K L, ZHANG M. Wireless capsule antennas [C]// Proceedings of the IAENG International MultiConference of Engineers and Computer Scientists (IMECS'13). Hong Kong, 2013: 726-729.

[19] KWAK S I, CHANG K H, YOON Y J. Ultra-wide band spiral shaped small antenna for the biomedical telemetry [C]// APMC. Suzhou, China, 2005: 4-7.

[20] LEE S H, CHANG K H, YOON Y J. A dual spiral antenna for wideband capsule endoscope system [C]// APMC. Bangkok, Thailand, 2007: 227-230.

[21] LEE S H, CHANG K H, KIM K J, YOON Y J. A conical spiral antenna for wideband capsule endoscope system [C]// AP-S. San Diego, US, 2008.

[22] LEE S H, YOON Y J. Fat arm spiral antenna for wideband capsule endoscope systems [C]// RWS. New Orleans, US, 2010: 579-582.

[23] LIM E G, WANG Z, CHEN J H, TILLO T, MAN K L. Study on the electromagnetic wave propagation of Wireless Capsule Endoscopy in human body [J]. Journal of Central South University of Technology, 2013, 20(10): 2650-2656.

(Edited by YANG Bing)

Foundation item: Project(BK20131183) supported by the Natural Science Foundation of Jiangsu Province, China; Projects(RDF-14-03-24, RDF-14-02-48) supported by Research Development Fund of Xi’an Jiaotong-Liverpool University, China

Received date: 2015-09-17; Accepted date: 2016-01-08

Corresponding author: Eng Gee Lim, Professor; E-mail: enggee.lim@xjtlu.edu.cn