Antennas, Wave Propagation, and Localization in Wireless Body Area Networks

Sammanfattning: A network of communicating wireless devices that are implantable, wearable or within close proximity of a human body is called wireless body area network (WBAN). The propagation channels for the devices in the WBAN are either through the body or over the body. This results in the attenuation and the absorption of electromagnetic waves radiated by the antenna of these devices due to the lossy tissues of the body. With a proper antenna and knowledge of the signal loss between the devices in the WBAN, a reliable wireless link can be designed. This thesis presents the investigations done for the antennas, wave propagation, and localization for various applications of these networks. The investigated applications are: (1) Binaural Hearing Aids (Paper I \& Paper II), (2) Sensor placed Around the Body (Paper III \& Paper IV), (3) Localization of Wireless Capsule Endoscope (Paper V), and (4) In-Mouth Devices. Binaural hearing aids communicate with each other for synchronization such as adjustment of volume or programing for the listening environment. In Paper I, antennas suitable in size and technical performance at 2.45 GHz for in-the-ear (ITE) and in-the-canal (ITC) placement of the hearing aids are presented. The ear-to-ear link loss found from the finite-difference-time-domain (FDTD) simulations for the ITE case was 48 dB and that for the ITC case was $92$~dB for the SAM head. The ITE case was further investigated on realistic heterogeneous phantoms of different age and head sizes. The link loss in the ITE case for an adult heterogeneous phantom was found to be $79$~dB. It was found that the absence of the pinna (outer ear) and the lossless shell under-estimates the link loss for the SAM phantom. Hence, a phantom with the lossy outer shell and the pinnas should be used for a proper estimation of the ear-to-ear link loss. In Paper II, an analytical model is presented for the ear-to-ear link loss based on the attenuation of the creeping wave over an elliptically modeled cross-section of the head. The model takes into account the dominant paths having most of the power of the creeping wave from the antenna in one ear to the antenna in the other ear and the effect of the pinnas. Simulations were done to validate the model using the ITE placement of the antenna at 2.45 GHz on heterogeneous phantoms of different age-groups and head sizes, showing a good agreement with the model. The effect of the pinnas was verified through measurements on a phantom where the pinnas fabricated by 3D-printing were included. In Paper III, an analytical model is developed for wireless propagation around the body based on the attenuation of creeping waves over an elliptical cross-section of the torso. The model includes the effect of the arms. It was verified through FDTD simulations on a numerical phantom with various arm positions. It was shown that it is critical to include the effect of the arms as their presence might result in extreme fading dips at some sensor positions. Further, a temporal variation in the power received by sensors placed around the torso was found when the arms moved while walking. This result is used in Paper IV to develop an approach to analyze the movements of the arms while walking using three wearable wireless devices. One of the devices is a transmitter placed at the back and the other two are symmetrically placed receivers at the side that record the power variation due to the arm movements. For such a placement of sensors, the variation in the receiver will be more or less symmetrical if the arm swing normally. However, a large degree of asymmetry in the arms swing will result in an asymmetrical variation in the received power by the two receivers. This was confirmed by simulations and measurements. Wireless capsule endoscopy (WCE) overcomes the problems of conventional endoscopy like not reaching the entire small-intestine. However, due to the lossy tissues, the localization of the capsule is a challenging task which is required in order to know the position of an abnormality captured in the endoscopy image. Paper V presents a method for the localization of an in-body RF source, as in WCE, based on microwave imaging. The electrical properties of the tissues and their distribution were found from microwave imaging at $403.5$~MHz. The method was applied on synthetic data obtained after addition of the white Gaussian noise to the simulated data of a simple circular phantom, and a realistic phantom for the 2D transverse magnetic polarization. The root-mean-square of the error distance between the various actual and estimated positions was found to be within $9$~mm for both the phantoms showing the capability of the algorithm to localize the capsule in the presence of noise with a good accuracy. An in-mouth device is a tongue controlled device used by the paraplegic or quadriplegic patients to control a wheelchair or type on a computer. Although the device is not surgically implanted but the performance of its antenna is influenced by the lossy tissues of the mouth or head in a similar way as that of an implant. In this work, antennas suitable in size and performance for such devices were investigated.