Radio and propagation for WBASNs
This section discusses the state-of-the-art for various existing channel models for WBASN applications. The first prototype platform of BoWI project is based on IEEE 802.15.4 ZigBeeTM infrastructure as will be discussed in the coming chapters, therefore the core work in this thesis is focused on unlicensed industrial, scientific and medical (ISM) band i.e., 2.40-2.4835 GHz. Moreover, there is also a future perspective of BoWI to utilize Ultra-wide band (UWB) i.e., 3.1-10.6 GHz in the next generation BoWI sensors due to the improved distance estimation accuracy of UWB. Therefore, state-of-the-art for UWB is also mentioned here.
Unfortunately, there is no generic channel model for WBASNs in contrast to wireless sensor networks. This is because of a number of limitations of WBASNs, some of which are as follows:
1. The most important consideration is the proximity of human body, being in the near-field of antennas comprising the wireless sensor nodes. This is because of high permittivity and conductivity of human body which causes significant losses and negatively impacts the antenna near-field distributions and hence, the radiation patterns and efficiency.
2. The propagation channels are time-varying because of the continuous body movements in contrast to typical WSNs which are subjected to stationary environments most of the time.
3. The channel is subject-specific and hence, can vary based on the body morphology e.g., gender, age, weight, height etc.
4. Body-centric channel is also sensitive to sensor placement and may need special requirements to minimize and detect placement error or calibration parameters.
5. The positioning of wireless nodes (or antennas) on the body such as their spacing or body proximity has significant impact on the channel. Antennas may be in line-of-sight (LoS) or non-line-of sight (NLoS) so can be shadowed or unshadowed by the body based on the subject body posture.
6. Surrounding environment would also influence the channel. Indoor and outdoor environments would offer different multipath profiles because of the surrounding objects and body motion will also introduce Doppler effect.
7. The channel models include individual antenna effects when proposed from measurements made with particular antennas (or sensors). This means that a specific channel model may not be applied with reliable accuracy for other antenna types (which are not used during its modeling). A more accurate way is to withdraw antenna characteristics from the channel model, a phenomenon named as antenna de-embedding. Once antennas are de-embedded, a channel may be valid with considerable accuracy for other types of antennas as well.
A number of channel models for WBASNs have been proposed in the literature based on different scenarios and constraints e.g., , - , and - . Both the static and dynamic channel models have been proposed so far. Static scenarios are those where there is no significant motion of the BAN subject, however, slight involuntary movements may occur e.g., breathing movements caused by the chest whereas dynamic scenarios include significant or forced body movements e.g., walking, running or routine activities or chores etc. The static channel models are mostly characterized by path loss whereas for dynamic channel models, number of other parameters also need special attention e.g., shadowing, power delay profile, Doppler spectrum etc.
The Task Group TG6 of IEEE802.15.6  is primarily concerned with modeling of wireless channels for WBASNs for medical and non-medical devices that can be placed inside body (implants) or on surface of human body. It has received number of contributions from different authors concerning different scenarios. A summary of various models can be found in .
Phantoms for modeling human body
Once the electrical characteristics of human body are known, the body can be modeled by physical or emulated structures called phantoms. Phantoms can emulate electrical characteristics of human body for simulation or measurement purposes at desired frequencies. This is also necessary in scenarios where in-vivo measurements cannot be taken. This makes phantoms ideal for medical research in areas such as X-ray, magnetic resonance imaging (MRI), and hyperthermia applications. Moreover, specific absorption rate (SAR) studies can also be conducted using phantoms which are not feasible with living subjects due to radiation dose safety concerns. Various safety standards specify the acceptable levels of radiation in terms of SAR such as those provided by the International Commission on Non-Ionizing Radiation Protection (ICNIRP)  and the Institute of Electrical and Electronics Engineers (IEEE) .
Physical phantoms are made from solid, liquid, or gel and are intended to make measurements in controlled laboratory environments. They can be classified on the basis of the tissue types they represent e.g., they can represent low-water content tissue, such as bones and fats, having low permittivity and low loss. They can also represent high-water content tissues such as brain, skin and muscles, which have higher permittivity and loss. Based on their final state of matter, they can be classified as solid (dry), semisolid (gel) and liquid phantoms.
Liquid phantoms are composed of a container filled with liquid having same electrical characteristics as the tissue in the human body, for the defined frequency range. The liquid is enclosed in a thin shell, usually made of fiberglass material with low relative permittivity and conductivity . Most recipes for the liquid contain sugar, diacetin or di-ethylene glycol butyl ether (DGBE) in different proportions to control permittivity of the solution, while salt (NaCl) is used to adjust conductivity  of the solution. Such phantoms do not represent human body accurately, since their internal structure is replaced with a homogeneous medium. Moreover, they do not allow measurement of SAR close to the surface of the body. However, these phantoms have the advantage of being easiest to fabricate. They are useful for experimental study of on- and off-body scenarios where the antennas are located outside the body perimeter and there is negligible field distribution inside the body due to the high medium attenuation e.g., at microwave frequencies. Such phantoms are not suitable for experimental study of in-body scenarios i.e., for implants. Liquid phantoms proposed by Ogawa et al.  and  are shown in Fig. 1.14.
In these phantoms, self-shaping coagulants are used to replace the outer shell of liquid phantoms. One of the popular gel phantom was developed by Guy , which is composed of water, sodium chloride, TX-150 (polyamide resin), and polyethylene powder. Ito et al.  have developed the self-shaping phantom based on Guy’s recipe adding sodium dehydroacetate and agar, used as preservative and coagulant to Guy’s recipe. These types of phantoms are suitable only for simulating high-water tissues such as muscle and brain but can adjust the electrical characteristic over a wide frequency range . The disadvantage of such phantoms is that the materials degrade over time, due to the loss of water and growth of fungi. An agar-based semisolid phantom -  is shown in Fig. 1.15.
Table of contents :
Chapter 1. Context and state-of-the-art
1.1. General Introduction, Wireless Body Area Sensor Networks
1.2. Introduction to BoWI project
1.3. Radio and propagation for WBASNs
1.3.1. On-body channel models
1.3.2. Off-body channel models
1.4. Electromagnetic characterization and modeling of human body
1.4.1. Electromagnetic characteristics of body tissues
1.4.2. Phantoms for modeling human body
22.214.171.124. Physical Phantoms
I. Liquid phantoms
II. Semisolid phantoms
III. Solid (dry) phantoms
126.96.36.199. Numerical Phantoms
I. Theoretical phantoms
II. Voxel phantoms
1.5. Antennas for WBASN applications
1.5.1. Antennas for ISM band and UWB
1.5.2. Antennas with material enhancements
1.5.3. Wearable/Textile Antennas
1.5.4. Adaptive or pattern-reconfigurable antennas
1.7. Thesis Structure
Chapter 2. Radio channel characterization for WBASNs using ultraminiaturized chip antennas
2.2. Zyggie prototype sensors and chip antennas
2.3. Characterization of chip antennas on an arbitrary substrate
2.4. Off-body diversity channel measurements
2.4.2. Measurement set up
2.4.3. Power delay profile and delay spread
2.4.4. Radio channel capacity
2.5. Distribution fitting for WBASN channel fading
2.5.2. Proposed Fitting Algorithm
2.5.3. Theoretical validation and accuracy
2.5.4. Fitting statistical study for off-body channels
2.5.5. Fitting statistical study for on-body channels
Chapter 3. Antenna design and interaction with human body at ISM frequencies
3.2. Printed monopole antenna
3.2.1. Antenna design
3.2.2. Antenna-body interaction
3.3. Inverted-F antenna
3.3.1. Free space performance
3.3.2. On-body performance
3.3.3. Specific Absorption Rate Computation
3.4. Dual-mode Patch Antenna
3.4.1. Characteristics mode theory
3.4.2. TM01 mode
188.8.131.52 Free space performance
184.108.40.206. Performance in proximity to body
Chapter 4. Miniaturized chip antenna design and numerical channel simulator
4.2. Meander Chip Antenna
4.2.1. Performance in free space
4.2.2. Performance in proximity to body
4.2.3. Prototype measurement
4.3. Numerical channel simulator
4.3.2. Validation of body models
4.3.3. Mesh convergence and accuracy
4.3.4. Significance of antenna-body spacing
4.3.5. Demonstration with few examples
4.3.6. Application for posture classification
Chapter 5. Pattern and polarization diversity antennas
5.2. Shorted ring patches based pattern diversity antenna
5.3. Dual-mode diversity antenna with improved cross-coupling
5.3.1. Antenna design
5.3.2. On-body performance
5.3.3. Experimental Characterization
5.3.4. Envelope correlation coefficient and diversity gain
5.4. Pattern and polarization diversity for on-body applications