Underground Wireless Sensor Network Communication Using Electromagnetic Waves Resonates at 2.5 GHz

A comprehensive study of electromagnetic waves underground propagation for a wireless sensor network is introduced in this paper. A mathemat ical model for path loss due to attenuation of electromagnetic waves propagating in different types of soil is given. Reflection from the air-soil interface as a function of distance between sensors and soil depth is also introduced. Three different types of soil, sandy, loamy and magnetite soil are studied. A high gain antenna is required to overcome the high value of path loss. A printed microstrip circular antenna is very common antenna used for underground wireless communication applications. A high gain microstrip circular antenna is designed and simulated using FEKO software. The antenna performance parameters studied in this paper are return loss, voltage standing wave ratio, input impedance and gain.


Introduction
A wireless sensor network (WSN) consists of sensors used to mon itor physical or environmental phenomena such as humid ity, temperature, sound, vibration, pressure or motion and cooperatively pass the data through the network of sensors to a main location [1]. As wireless network sensors become smaller in dimension and cheaper, researchers are deploying them in environments that are unconventional for electro magnetic signaling [2]. One of those applications for wireless sensor network is underground wireless communicat ion to monitor soil properties and then transmit the collected data to a node on the surface [3]- [8].
Wireless underground sensor network has a lot of applications including: environ mental monitoring, infrastructure monitoring, location determination and border patrol and security monitoring [9,10].
Wireless communication using electromagnetic signals suffers from h igh attenuation in soil, due to the absorption of the signal. So, the underground wireless communication can be characterized by high signal loss due to absorption and mu lti-path losses caused by soil molecules such as reflection and refraction [11]. The ma in advantages of using electro magnetic waves instead of sound are: first, electro magnetic waves reduce the latency due to faster propagation. Second, electro magnetic waves give a high data

Related Work
There has been some work focusing on electromagnetic waves propagation through soil and water. N. Channmwe at al. introduced the empirical attenuation and relative permittiv ity values for different materials including soil at 300-700 MHz frequency range [14]. Reflection and refract ion fro m the underground-air interface are also introduced in this paper. The path loss due to material absorption in underground propagation is presented by N. Peplinski. Path loss as a function of volumetric water content 5%, 10%, 15%, 20% and 25% for a frequency range 1.3-3 GHz is introduced in this paper [15].
A bit error rate and path loss as a function in frequency and sensor depth for different volu metric water content are introduced by I. Akyild iz at el. [16]. There are several challenges in the design of wireless underground communicat ion network such as network topology, reliability, localization and network architecture design.
The communicat ion through soil is represented as an electro magnetic wave transfer through the transmission line by T. Weldon and A. Rathore [17]. The operating frequency range is 1-2 GHz. In [18], the authors show that the soil composition has significant effects on the ground penetrating radar.

Effect of Soil Properties on Underground Communication
The soil properties have a great effect on the underground communicat ion using electromagnetic waves. Dielectric constant is most important parameter that affects the propagation in underground communication. Water content, density, particle size and temperature are also affecting the communicat ion.

Water Content
Signal loss through a given type of soil is depending on water concentration in soil. Increasing the water content of a soil makes the channel mo re power loss. As water content raise fro m dry the 13% volu metric, the losses per meter increased by 137 dB at 1 GHz frequency [17].

Density
The path loss increases as the soil density is increasing. Also, the signal attenuation increases with increasing the soil density.

Particle Size
There are three d ifferent types of soils accord ing to particle size. Three major co mponents are given in [18] as sand, silt and clay. Sandy soil is the smallest particle in size and gives a lower losses and clay soil gives the highest losses for its large particle size [19].

Temperature
Dielectric properties are function of temperature, so the increasing in temperature leads to increase the signal attenuation in underground communication.

Antenna Background
Underground communication needs a very efficient antenna for wireless sensor network co mmunication. This antenna must meet a number of requirements for underground communication to overcome the high value of path loss due. This kind of antenna must has a high gain, above 10 dB, and should be small in d imension so that it can be fitted on the sensor surfaces. The most common antenna used for underground communication using electro magnetic signals is circular microstrip antenna [2].

Background
A circu lar microstrip antenna is a popular for frequency range for Ultra High Frequency (UHF), fro m 300 MHz to 3 GHz, up to the millimeter wave range, fro m 30 GHz to 300 GHz, and has also found application in arrays. The circular antenna performance is not sensitive to small parameter variations, imp roving robustness to manufacturing tolerances. While the circular microstrip antenna provides reasonable wide-band performance, this is not a high performance antenna [20,21].

Physical Descripti on
The circular microstrip antenna is easy to construct and can be very robust, but can become restrictively large at low frequencies. This antenna is an electrically large antenna with a one wavelength circu mference. The gain is stable over a substantial band; however, the input impedance shows a large variation [22,23].

Underground Signal Propagation
The signal propagation in soil depends on the path loss in soil. Received power as a function of transmitted signal, path loss and antenna gain at the receiver end is given fro m Friis equation as shown in Equation 1 [24].
where P t is the transmit power, G r and G t are the gains of the receiver and transmitter antenna, L Pathloss is the path loss in soil.
The path loss is shown in Equation 2 [25].
0 is the path loss in air and given by: where d is the distance between transmitter and receiver in meter, f is the operating frequency in Hertz and c is the velocity of light in air in meter per second.
( ) is the path loss due to changing in mediu m and given by [26]: where λ 0 is the signal wavelength in air and calculated (λ 0 =c/f) and λ is the wave factor and given by (λ=2π/β) and β is the phase shifting constant and calculated as shown in Using Electromagnetic Waves Resonates at 2.5 GHz

Reflection from Ground Interfaces
The reflection fro m the surface and bottom depends on reflection coefficient at the interface between soil and air. The reflection coefficient is given by Equation 8 [27]. Г = 2 2 − 1 1 where ρ 1 where r is the reflected path length, | Г | and are the amp litude and phase of the reflection coefficient respectively and Δ(r) is the difference between r and d.
where r can be calculated as follow: where d is the distance between two sensors, H is the distance between surface and the sensor and r is the distance between the sensor and the reflection point.

Results
The effect of frequency on the path loss for different values of distance between sensors using different soil types is illustrated in the following sections. Nine different soil types are shown in Tab le 1 according to the water concentration in soil.

Path Loss Calculation
The total path loss due to communication between sensors without reflection loss is shown in the ne xt sections.

Sandy Soil
The sandy soil is the normal pure soil. The Particle density, which is the weight per unit volu me of the solid portion of soil, of normal soils is 2.65 g/cm3. The Bulk density, the oven dry weight of a unit volu me o f soil inclusive of pore spaces, of a soil is always smaller than its particle density and is about 1.6 g/cm 3 and the pore space is 40% [28].
The sandy soils with three different concentration of water are studied in this section. The path loss for the dry sand soil with relative permittivity ɛ`=2.55 and tangent loss is tan(δ)=0.0062, sandy soil 3.88% water with ɛ`=4.4 and tan(δ)=0.046 and sandy soil 18.8% water with ɛ`=20 and tan(δ)=0.13 at 2.5 GHz are g iven.  The effect of frequency on the path loss at distance between two sensors 1 m is illustrated in Figure 2. As clearly shown in the figure, as the frequency increases the path loss is also increases. The path loss is increased as the water concentration is increased due to increasing in the dielectric constant of the sandy soil. In Figure 3, the effect of d istance on a path loss is illustrated for different values of water concentration at frequency 2.5 GHz.

. Loamy Soil
The Bulk density of the Loa my soil is 1.4 g/cm 3 and the pore space is 47% which means the Loamy soil become finer in texture [29].
The loa my soils with three d ifferent concentrations of water are introduced in this section. The path loss for the dry loa my soil with relat ive permitt ivity ɛ`=2.44 and tangent loss is tan(δ)=0.0011, loa my soil 2.2% water with ɛ`=3.5 and tan(δ)=0.04 and loamy soil 13.77% water with ɛ`=20 and tan(δ)=0.12 at 2.5 GHz are introduced.
The frequency as a function on the path loss for different water concentration for loamy soil at distance 1 m separation between sensors is illustrated in Figure 4. Increasing water concentration leads to increasing path loss. In Figure 5, the effect of distance on a path loss is illustrated at frequencies 2.5 GHz for loa my soil.

Magnetite So il
A sandy soil with a large amount of heavy minerals such as magnetite is called a magnetite soil. The particle density is getting higher for magnetite soil [30]. The magnetite soil is a pure soil contains iron o xide part icles [31].
The magnetite soils with three different concentration of water are introduced in this section. The path loss for the dry magnetite soil with relative permittivity ɛ`=1.05 and tangent loss is tan(δ)=0.029, magnetite soil 4.8% water with ɛ`=8.3 and tan(δ)=0.22 and magnetite soil 11% water with ɛ`=30 and tan(δ)=0.32 at 2.5 GHz are introduced.
The frequency as a function on the path loss for different water concentration for magnetite soil at distance 1 m separation between sensors is illustrated in Figure 6. Increasing water concentration leads to increasing path loss. In Figure 7, the effect of distance on a path loss is illustrated at frequencies 2.5 GHz for magnetite soil.

Reflecti on Calculation
An extra loss due to reflection is obtained. The reflection fro m soil air interface is studied in this section. So me approximations are assumed here to simplify the simu lation as in Figure 1 as fo llo ws: • The sensors are in the middle of the sand height, H 1 = H 2 =H.
The path loss due to reflect ion fro m soil surface interface is calculated as a function of sensor depth H at distance between sensors 1 m and 2.5 GHz frequency for three different soils. As shown in Figure 8, the effect of water concentration on the reflection loss is very high for the dry soil. As the water concentration and sensor depth increased, the path loss decreased to be negligible at depth 5 m. The same conclusion is obtained in Figure 9. Path loss due to reflection as a function of depth at 2.5 GHz and distance 1 m is illustrated in Figure 9 for loamy soil. Reflection for magnetite soil at different water concentration after 3 m depth is almost vanishing as shown in Figure 10.

Antenna Design
Circular microstrip antenna operates at 2.5 GHz design is shown in Figure 11. FEKO software is used to design and simu late this antenna. The dimensions of the designed antenna are as follow: antenna patch diameter is 30.36 mm, substrate height is 1.5 mm, substrate length is 75.9 mm, substrate width is 75.9 mm and the dielectric constant is 4.35.  Figure 13. Return loss value is -25 dB at 2.5 GHz wh ich is very efficient for using in underground use, should be less than -10 d B in most of underground applications. Figure 14 shows the voltage standing wave ratio for the designed antenna. The VSWR value is 1.1 at 2.5 GHz which is very efficient in manufacture process of the microstrip circular antenna. Real part of input impedance is shown in Figure 15; the real value of input impedance is almost 50 oh m at 2.5 GHz which match the transmission line impedance. The maximu m gain is obtained at 2.5 GHz at the radiation angle 00 for the designed antenna as shown in Figure 15.

Conclusions
A mathemat ical model for path loss due to attenuation of electro magnetic waves propagates in different types of soil resonates at 2.5 GHz is introduced in this paper. The reflection of electro magnetic waves at the air-soil interface is given as a function of sensor depth in soil and distance between sensors. Three different types of soil are given in this paper, sandy, loa my and magnetite soil. For lower permittiv ity, the path loss is a low value. The higher distance between sensors leads to higher losses is produced.
The reflection fro m air-soil interface is negligible in case of deep soil, after 3 m, except for dry sandy soil and dry loa my soil.
A high gain microstrip circu lar antenna is designed and simu lated using FEKO software. Return loss, voltage standing wave ratio, real part o f input impedance and gain is also given in this paper. S11 of the designed antenna is -25 dB at 2.5 GHz, which is very efficient value for underground wireless commun ication to overco me the high path loss due attenuation.