INTRODUCTION



Confined Intra-Arm Communication and Medical Applications – Extended AbstractMSR-TR-2012-74Trang ThaiGeorgia Institute of Technology85 5th St., NWAtlanta, GA 30308 U.S.A.001-404-944-0845trang.thai@gatech.eduGerald DeJeanMicrosoft ResearchOne Microsoft WayRedmond, WA 98052 U.S.A.001-425-722-6400dejean@Ran Gilad-BachrachMicrosoft ResearchOne Microsoft WayRedmond, WA 98052 U.S.A.001-425-706-7901rang@ABSTRACTPersonal area networks are enablers for many new medical applications. In this work, we present an implementation of such a network through guided wave on body communication channel. This method allows the creation of high bandwidth communication channels which are confined to the body and improve on previous technologies in terms of privacy and resilient to interference and bandwidth. The technology proposed can also be used for other applications such as tracking infections and especially Nosocomial infections. Nosocomial infections (hospital-acquired infections) are believed to be linked to the death of around 100,000 patients each year in the U.S. only. Improving this situation requires monitoring the interactions between patients, staff members and objects in the hospital. The technology proposed here allows the detection of a hand shake between people and the interaction with other objects and thus registering them for analysis of the root cause of an infection. Categories and Subject DescriptorsB.4.3 [Interconnections (Subsystems)]: Physical structures J.3 [Life and Medical Sciences]: healthGeneral TermsMeasurement, Performance, Design, Experimentation, Human FactorsKeywordsPersonal area network, antenna design, infection tracking, sensor networkINTRODUCTIONMany evolving health care application use a network of sensors and actuators spread on the subjects’ body. Personal area networks (PANs) allow sensors and actuators to communicate with each other and with other components such as gateways and computing providers [1]. In this work, we discuss a special type of network in which the communication channel is the body itself, as opposed to the air as in Bluetooth, IEEE 802.15.4 and similar technologies. Traditional, air-based networks, suffer from several limitations. The gain of the network is severely affected by the posture of the subject [2]. They are subject to interference and require sophisticated privacy preserving protocols in the logical layers. On-body channels have several advantages [3]. They are confined, by design, to the body and hence are subject to less interference and less privacy and security concerns. They also enable a set of applications such as transactions approved by handshakes [4] and more.In this work, we present a novel body-wave method that improves on previous on-body communication solutions. The main observation is that the skin is conductive and the arm is cylindrical. Therefore, we design a prototype arm band that transforms the arm into a modified coaxial line. In a nut-shell, the antenna uses the arm as the inner lid in a coaxial cable. The arm band itself injects the signal to the arm and acts as an excitation source and part of the coaxial channel. In the modified form, the channel is a leaky coaxial cable on which the signal is highly confined. Note that coaxial cables are the most shielded transmission lines for high frequency signals. The resulting channel works at higher frequencies then previous solutions [3, 4] and thus, is less likely to interfere with natural body signals. It provides an order of magnitude higher bandwidth and is much more confined to the arm.The proposed solution allows implementation of many applications. For example, a temperature sensor placed on the arm pit can communicate with a wrist watch over this channel. If two people wearing the arm band shake hands, a communication channel is established which allows for applications such as money transfer, business card sharing or signing agreements [3].One of the interesting applications of this technology in the health domain is tracking the propagation of infections, especially in the hospital environment. Nosocomial infections (also known as hospital acquired infections or HAI) are a major cause for death and other complications. For example, around 20% of patients who go through a bone marrow transplant die due to infections and bacteria [5]. According to the report of the International Nosocomial Infection Control Consortium [6], around 7.6% of the patients in intensive care units (ICUs) suffer from HAI. They also report on a significant increase in the mortality rate in patients with HAI. For example, the crude mortality rate among patients without HAI is around 14:4% while the crude mortality rate for patients with HAI ranges from 18:5% to 42:7% depending on the type of infection [6] (Table 12 therein). Similar affects have been measured in infants [6] (Table 13 therein). Moreover, acquiring an infection increases the number of days a patient stays for treatment and those increase costs.To restrict the risk of antibiotic resistance, it is extremely important to restrict the use of anti-infectives [6]. Thus, it has been found that targeted programs for improving hygiene are very efficient in reducing HAI rates without increasing the risks of antibiotic resistance [7, 8]. These studies have shown that the problem of sustaining high degree hygiene levels is not just a matter of education. It turns out that among the hospital staff, the worst hygiene levels are maintained by physician. This is despite the fact that they are well educated and know about the significance of it, as found by Ignaz Phillip Semmelweis in the 19th century [9]. Therefore, continuous targeted performance feedback should be used to complement the educational tools [6]. The communication channel described in this work can be used for tracking the in hospital interactions such that once an infection is discovered it will be possible to trace back its root cause, whether it is a human or an object that was not treated properly. The focus of this paper is the introduction of our technology, especially the design of the arm band and the physics behind its capability. As discussed above, the motivation for the development of this technology stems from its medical applications. Unfortunately, at this stage we have not had the opportunity to conduct clinical trials.RELATED WORKOne of the biggest challenges of using the human arm as a medium for signal transmission is properly injecting a strong signal to the arm for propagation. Past research has extensively studied some techniques that can be used to excite signals into the human arm. Most of these methods focus on placing transmit and receive electrodes on the biological entity for transmitting and receiving the signal electrostatically [1, 4, 10]. In these cases, the communication is facilitated through electric field between two grounded electrodes (Fig. 1). Most of these studies have been performed at frequency ranges around 10 MHz and have transmission gains in the neighborhood of -30 to -25 dB. These are not guided waves but rather radiated waves or leaky waves, which are not to be confused with leaky channels.In addition, there was a brief explanation of a method that treats the human body as a waveguide (Fig. 2) with high frequency electromagnetic waves generated at a terminal propagating through the body that is captured by another terminal at a finite distance away [3]. However, the concept of utilizing the arm into part of the coaxial transmission line topology introduces a new mechanism of transmission as well as a channel platform with significantly altered signal distribution and signal confinement. To the authors’ knowledge, this concept has never been observed, and at the same time there have not been much results published based on body channel either. The objective of this method is that using the human arm as the inner conductor of the coaxial waveguide will allow the electromagnetic energy to be confined Figure 1. Static electric field communication through electrodes. Figure 2. Communication through wave propagation.mostly on the skin [11]. The conductivity of the skin medium is much higher than the fat layer residing underneath it. In this paper, we present a method of coupling an electromagnetic signal to the human arm through the use of a coaxial transmission line where the human arm acts as an inner conducting medium allowing an injected signal to propagate tangential to the outer circumference of the arm. The human arm is loaded with an open-ended coaxial transmission line arm band that is worn on the arm. A voltage signal is excited between the outer and inner conductors of an accompanying arm band. The inner conductor of this arm band is in direct contact with the arm, and the signal from the inner conductor is coupled to the arm for transmission. As will be seen later, a direct transmission line mode will be present on the arm in the 400-500 MHz range allowing for a deliberate means of wireless on skin communications. This presents a great opportunity for a myriad of applications such as health monitoring and tracking of infections in a hospital environment. ARM BAND DESIGNThe arm band is designed as an open-ended circular coaxial cuff that is placed around a human arm. The basic components of a standard coaxial cable are an outer metallic shield, a center core metallic inner conductor, and a dielectric insulator between the outer shield and inner conductor (Fig. 3). Electric field lines propagate in the dielectric medium (including air) between the conductors. Fig. 4 shows an illustration of this modified arm band used for the human arm. The inner conductor is in direct contact with the arm. Its diameter is 5 cm, while that of the outer conductor is 6 cm. Therefore, when this band is worn around the arm, a coaxial mode is initially excited between the inner and outer conductors. An illustration of the arm band around a cylinder modeled as a human arm is displayed in Fig. 5. The length of the inner conductor is designed to be shorter than that of the outer conductor. In turn, the conductive properties of the human arm replace the metallic inner conductor and become the new inner conductor for the coaxial transmission as the electric fields propagate down the length of the arm. Figure 3. Communication through wave propagation. Figure 4. Model of arm band.Then, at the termination of the outer conductor, the coaxial line is open and two types of electromagnetic phenomena occur at different frequencies. Within the frequency range of approximately 400-500 MHz, the electric fields at the open ended termination of the outer conductor stay closely confined to the conductive arm as shown in Fig. 6. This facilitates non-TEM (transverse electromagnetic) transmission line propagation of fields that begin to gradually lose intensity with the increasing length of the human arm. On the other hand, at the higher frequency of 1.5 GHz, the channel appears like a dielectric waveguide and the electric fields are distributed mostly over the inner region of the arm (the fat layer) rather than concentrated at the skin as in the case of 400 MHz signals as shown in Fig. 7 [12]. Notice that the field at 1.5 GHz is present at the center of the arm’s cross section, while the field at 400 MHz is scarcely present inside the arm [Fig. 6]. The electric fields terminated at the open ended termination are radiated into the air that generates an “antenna-like” mode. This mode offers an opportunity to communicate to base stations and other mobile devices and opens the door to several applications. Using the human arm as an inner conductive medium of a coaxial transmission allows for multiple modes of communication. Figure 5. Model of human arm loaded with arm bands. Figure 6. Electric fields at 400 MHz as probe plane travels from the source starting at ‘a’ and moving to ‘c’. Figure 7. Electric fields at 1.5 GHz as probe plane travels from the source starting at ‘a’ and moving to ‘c’.HUMAN ARM SIMULATION SETUPTo get an idea of the relative received power of one arm band in receive mode when another arm band transmits a signal, it is necessary to examine the path loss between the two devices. The path loss is a measurement of the amount of power that is loss when one device transmits a signal to a second receiving device. The path loss typically does not take into account any power losses due to imperfect conductors or lossy dielectric mediums. To examine the path loss, the authors performed simulations using two electromagnetic software packages: Microstripes and Microwave Studio (both developed by CST). The setup for this simulation consists of a cylinder modeled as a human arm that has a diameter of 5 cm. The biological materials used in the simulation are dry skin, fat, and muscle. The most inner material of the arm model is muscle followed by fat, and finally skin as the outermost layer. Table 1 illustrates the thicknesses and the electrical parameters of each of these biological materials. In simulation, these materials were treated as 2nd order dispersive Debye materials. A dispersive material is one in which the electrical properties change with frequency. The two arm bands are in direct contact with the arm and physically separated by a finite distance denoted L. In addition, the arm bands were excited MaterialInner/Outer DiameterDielectric ConstantMuscleNA/19mm472 @ 2.3 MHz, 1.37 @ 1.15 GHzFat19mm/24mm9282 @ 1 MHz, 8.75 @ 712 MHzDry Skin24mm/25mm24.71 @ 356 MHz, 1.6 @ 4.4 GHzin simulation by a linear 20 ohm reference discrete port between the inner and outer conductors. The figure of merit in the simulation in the S21 scattering parameter which measures the voltage ratio of signal received from one arm band in receive mode (called port 2) when the other arm band is in transmit mode (port 1): S21=V2-V1+ ,where V is the voltage (proportional to the square root of power), numeral 1 is one arm band, numeral 2 is another arm band, ‘+’ means transmit mode, and ‘-’ means receive mode. This parameter is used to represent the path loss between the two devices. Three simulations were performed: 1.) a S21 measurement between the two arm bands when L=40 cm; 2.) a S21 measurement between the two arm bands when L=30 cm; and 3.) a S21 measurement between the two arm bands when L=40 cm, but the arm-modeled cylinder has been split with a 5 cm gap. This signifies a scenario of two separate arms with one arm band being worn on each arm. Fig. 8 displays a plot of this S21 measurement versus frequency. In this plot, it can be seen that the S21 experiences maximum voltage ratios between 375 – 526 MHz for the three simulations. This frequency range corresponds to the range of transmission line signal propagation. As expected, the closer the arm bands are to each other without discontinuity, the higher the S21 parameter (meaning the lower the path loss). A low path loss symbolizes that a large amount of transmitted signal from one device has been received by a second device assuming only two devices are present. The presence of the gap in scenario #3 results in the lowest S21 (or the highest path loss). This is understandable since the current through the arm is discontinued at the gap. Since the gap is not very large, the electromagnetic fields at the gap are responsible for the magnitude of signal in this scenario. Figure 8. S21 measurement of path loss simulation for modeling a human arm loaded with arm bands.It is also important to examine the S21 response at the frequency range around 1.5 GHz. Here, the S21 magnitude is the same for all three scenarios. This is because the mechanism of signal propagation at this frequency is dominantly due to radiation. Therefore, the presence of a gap does not affect the transmission. Also, a plot of S11 is shown in Fig. 8 which signifies the ratio of signal received from a device to signal transmitted from the same device. This result is exactly the same for all three scenarios. Typically in antenna design, a S11 minimum below -6 dB for a single device signifies a radiation mode. Since there are two devices, one needs to analyze the S11 and the S21 response to determine if a radiation mode exists. Although the minimum dip in the 1.5 GHz range does not meet typical standards, combining this response with an S21 response lower than -40 dB allows one to logically assume that a significant amount of energy is radiated into the atmosphere.UNDERSTANDING THROUGH THE EXAMINING ELECTRIC FIELDSTo supplement the understanding of the transmission line mode (at around 400 MHz) and the radiation mode (at around 1.5 GHz) for this setup, the electric fields are displayed in Figs. 6,7 and 9. Fig. 6 shows the electric fields at 400 MHz as a probe plane travels down the length of the arm model starting at the source of the arm band where the field strength is the strongest. As this probe plane travels down the arm away from the arm band, a small quantity of electric fields continue to propagate tightly to the arm model. The intensity of electric fields is very small radially away from the arm model throughout the propagation. This behavior suggests that transmission line properties are indeed exhibited around 400 MHz. Note that in the simulation depicted in Fig. 7a, in the cross section plane located along the arm band, the field drops from -30 dB at the outer surface of the arm band to -50dB at 8mm away from this surface (in air). Compared to the field distribution generated by method reported in [4], the field drops about 10 dB between from the electrode surface and approximately 2 cm away into the air. Therefore, our on-skin channel platform presented here shows to produce highly confined signals allowing extra security and privacy at the physical layer of a communication network. Fig. 7 depicts a different type of wave propagation at 1.5 GHz. In this figure, a strong intensity of electric fields is present radially away from the arm model. This intensity remains strong radially away from the model as the probe plane travels down the arm. It is concluded that the signal propagation at this frequency is due to radiation. Finally, Fig. 9 shows the electric fields when a 5 cm gap is inserted into the model. At 400 MHz, the signal is tightly confined to the model without radial propagation, but at 1.5 GHz, the signal is radiated away from the arm regardless of the existence of a gap. Therefore, the transmission line properties at 400 MHz and the radiation properties at 1.5 GHz are maintained regardless of any discontinuities in the model.MEASUREMENT SETUPTo validate the simulations of the transmission line on-body propagation of signals, we designed prototype arm bands to be worn by actual human arms. Prototypes of the arm band are illustrated in Fig. 10. The arm band consists of circular copper strips – an inner strip connected to the arm and an outer strip to serve as the ground – separated by styrofoam cut into strips to allow for bending of the arm band. The copper strips have adhesive on one side to adhere to the styrofoam. Since most measurement equipment that is used to transmit signals are 50 ohm systems, a quarter-wave transformer were included to the Figure 9. Electric fields at 400 MHz and 1.5 GHz as probe plane travels from the source starting at ‘a’ and moving to ‘c’ for human arm model with a loaded gap.physical design to transform the 20 ohm used in simulation to 50 ohms. The quarter-wave transformer is a microstrip line that has an electrical length of λ/4 where λ is the wavelength = the speed of light in free space divided by the frequency. The microstrip line was soldered to the inner conductor of the arm band while the backside ground of the transformer was soldered to the outer conductor of the arm band. An HP 8510 vector network analyzer was utilized to send signals to the arm bands. Coaxial connectors were used as an interface between the network analyzer and the arm band. Since these are the beginning stages of this research, we are only interested at this time in setting up an experiment to justify that a signal can be obtained from a human arm when the arm bands are worn. To complete this experiment, two of the authors, Trang and Gerald, were used, and each of us wore an arm band that was excited with a network analyzer. With both arm bands excited, a coaxial probe is used to capture the change in power along the length of the arm. This coaxial probe is connected to the input port of a spectrum analyzer to display the change. Since it is possible for the probe to pick up ambient signals that are identified as noise, it is important to remember that the relative change of the signal when the probe touches the arm is the parameter that is being investigated. The entire measurement setup is shown in Fig. 11.MEASUREMENT RESULTSSince the coaxial probe is merely the short signal line of a coaxial connector, the length of the signal line is too short to capture anyFigure 10. Prototype arm band and coaxial probe. VNASpectrum AnalyzerVNASpectrum AnalyzerFig. 11. Measurement systemradiative signals. Therefore, we limit the experiments to the transmission line characteristic of the communication channel. The measurements that were performed include two coaxial probe conditions: poke and non-poke. The poke condition is active when the probe is normal to the arm and in direct contact with the arm as the arm bands are transmitting and receiving signals. The non-poke condition constitutes the probe being tangent to the arm at a radial distance of approximately 5 mm away without actually touching the arm. Fig. 12 shows the received power from the coaxial probe for poke and non-poke contact at two distances (15 cm and 5 mm away along the arm axis) from each individual’s arm band when there is a 5 cm spacing between our arms. This condition is similar to scenario #3 used in simulations. In these plots, it is noted that the poke condition applied to both individuals records a larger received power than the non-poke condition on both individuals. This can be seen across the frequency band, but the high received power at the 1.5 GHz radiative band when the poke condition is enforced can be attributed to transmission as well as near-field radiation from the arm as a current-carrying source. The difference in magnitude between Trang’s poke condition and Gerald’s poke condition can be attributed to many factors including differences in our body’s electrical current and ambient sources of noise. Notice that despite the noisy signal, the characteristics of the channel is shown clear in the case of 15 cm away in which the difference of the power level between poke and non-poke condition is the largest across the band.Fig. 13 illustrates the received power from the coaxial probe for poke and non-poke contact but for this measurement, only one arm band was worn at a time. Additionally, Trang and Gerald were shaking hands, thus establishing a continuous connection and transfer of body current. This plot gives a good understanding of the transmission line properties of the communications channel. Again, one can see that the poke conditions result in a higher received power that the non-poke condition. A measurement was also recorded for a poke condition on Gerald’s arm when Trang was wearing the arm band while shaking hands and vice versa. Here, it is important to notice that the received power is much larger in the 400 MHz range than in the 1.5 GHz range. Compared the non-poke to poke on the non-equipped arm in the hand shake, the high power received in the 400 MHz range successfully validates the on-skin channel with its confinement characteristic. This characteristic can be attributed to the decreasing signal strength at the higher frequency (non-guided waves) that does not allow for a strong radiation. Conversely, a relatively strong signal is received from the tightly confined transmission line mode at the lower frequency range. Figure 12. Received power from coaxial probe when Trang and Gerald are not in contact with each other. Figure 13. Received power from coaxial probe when Trang and Gerald are shaking hands with one arm band active.Finally, Fig. 14 shows the received power from the coaxial probe for poke and non-poke contact when Trang and Gerald are both wearing the arm bands while shaking hands as illustrated in Fig. 15. In this measurement, the poke condition occurs at the center of Figure 14. Received power from coaxial probe when Trang and Gerald are shaking hands with both arm bands active.Figure 15. Setup of hand shake condition in which both worn arm bands are excited by a VNA while the measuring probe is connected to a spectrum analyzer for signal level detection.the connection of the hands. Similar to the other cases, the poke conditions exhibit a larger received power than the non-poke conditions, which strongly suggests that the signal at 400 MHz range takes the guided form and well confined to the surface so that its propagation on the surface is the strongest. The difference in received power between poking Trang’s arm and poking Gerald’s arm can be attributed to some unknown factors that we did not examine in these experiments. More rigorous studies of using the human body, particularly the arm, will be intensely explored in future work, but one common trend that Figs. 12-14 exhibit is the increase in received power from contact with a human when an arm band is transmitting.CONCLUSIONSIn this work, we present a novel method to create on body communication channel. Using simulations, we showed that this channel has high capacity, compared to previous technologies and is more confined to the body. Thus, it has higher throughput and better privacy and resilient features. We have validated the predictions of the simulation using a prototype. We have discussed the application of this communication channel, especially in the medical domain. It allows the creation of on-body communication network between sensors, actuators and other components. Moreover, it may be used as a sensing device since it can register the physical interactions between people and the interactions with certain objects. We suggest that this may play a key role in tracking infections, especially in the hospital environment.We would like to further explore the properties of the proposed channel. While the simulations provide encouraging results which we validated with a prototype that we have built, a more comprehensive study using the prototype is needed. Moreover, we would like to test the effectiveness of the approach in a real medical scenario.ACKNOWLEDGMENTSThis research was conducted while Trang Thai was visiting Microsoft Research. The authors would like to thank Mike Sinclair for his involvement in valuable discussions that led to the measurement results of this paper.REFERENCESZimmerman, T. G. (1995). Personal Area Networks (PAN): Near-Field Intra-Body Communication. IBM Systems Journal.Alomainy, A., Owadally, A., Hao, Y., Parini, C., Nechayev, Y., Constantinou, C., et al. (2004). Body-Centric WLANs for Future Wearable Computers. International Workshop on Wearable and Implantable Body Sensor Network.Hachisuka, K., Nakata, A., Takeda, T., Terauchi, Y., Shiba, K., Sasaki, K., Hosaka, H., Itao, K. (2003). Development and Performance analysis of an intra-body communication device. International Conference on Transducers, Solid-State Sensors, Actuators and Microsystems. IEEE.Fujii, K., Takahashi, M. and Ito, K. (2007). Electric Field Distributions of Wearable Devices Using the Human Body as a Transmission Channel. IEEE Trans. on Antennas and Propagation.Gooley, T. A., Chien, J. W., Pergam, S. A., Hingorani, S., Sorror, M. L., Boeckh, M., et al. (2010). Reduced Mortality after Allogeneic Hematopoietic-Cell Transplantation. The new england journal of medicine.Rosenthal, V. D., Maki, D. G., Jamulitrat, S., Medeiros, E. A., Todi, S. K., Gomez, D. Y., et al. (2010). International Nosocomial Infection Control Consortium (INICC) report, data summary for 2003-2008. American Journal of Infection Control.Tibballs, J. (1996). Teaching hospital medical staff to handwash. The medical journal of Australia.Rosenthal, V. D., Guzman, S. and Safdar, N. (2004). Effects of education and performance feedback on rates of catheter-associated urinary tract infection in intensive care units in Argentina. Infection control and hospital epidemiology.Wyklicky, H. and Skopec, M. (1983). Ignaz Philipp Semmelweis, the prophet of bacteriology. Infection Control.Zimmerman, T., Smith, J. R., Paradiso, J., Allport, D., Gershenfeld, N. (1995). Applying Electric Field Sensing to Human-Computer Interfaces. Conference on Human Factors in Computing Systems, ACM Press. Hoffmann, J., Leuchtmann, P., Ruefenacht, J., Hafner, C. (2009 ). Propagation constant of a coaxial transmission line with rough surfaces. IEEE Trans. On Microwave Theory and Techniques.Buck, J. (1995). Fundamentals of Optical Fibers. Wiley-Interscience. New York, NY ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download