sensors Article

Implementation and Analysis of ISM 2.4 GHz Wireless Sensor Network Systems in Judo Training Venues Peio Lopez-Iturri 1 , Erik Aguirre 1 , Leyre Azpilicueta 2 , José Javier Astrain 3 , Jesús Villadangos 3 and Francisco Falcone 1, * 1 2 3

*

Electrical and Electronic Engineering Department, Public University of Navarre, Pamplona 31006, Spain; [email protected] (P.L.-I.); [email protected] (E.A.) School of Engineering and Sciences, Tecnologico de Monterrey, Campus Monterrey, Monterrey, NL 64849, Mexico; [email protected] Mathematical Engineering and Computer Science Department, Institute for Smart Cities, Public University of Navarre, Pamplona 31006, Spain; [email protected] (J.J.A.); [email protected] (J.V.) Correspondence: [email protected]; Tel.: +34-948-169-741

Academic Editor: Stefano Mariani Received: 31 March 2016; Accepted: 2 August 2016; Published: 6 August 2016

Abstract: In this work, the performance of ISM 2.4 GHz Wireless Sensor Networks (WSNs) deployed in judo training venues is analyzed. Judo is a very popular martial art, which is practiced by thousands of people not only at the competition level, but also as part of physical education programs at different school levels. There is a great variety of judo training venues, and each one has specific morphological aspects, making them unique scenarios in terms of radio propagation due to the presence of furniture, columns, equipment and the presence of human beings, which is a major issue as the person density within this kind of scenarios could be high. Another key aspect is the electromagnetic interference created by other wireless systems, such as WiFi or other WSNs, which make the radio planning a complex task in terms of coexistence. In order to analyze the impact of these features on the radio propagation and the performance of WSNs, an in-house developed 3D ray launching algorithm has been used. The obtained simulation results have been validated with a measurement campaign carried out in the sport facilities of the Public University of Navarre. The analysis is completed with the inclusion of an application designed to monitor biological constants of judokas, aimed to improve their training procedures. The application, that allows the simultaneous monitoring of multiple judokas (collective workouts) minimizing the efforts of the coach and medical supervisor, is based on commercial off-the-shelf products. The presented assessment of the presence of interfering wireless systems and the presence of human beings within judo training venues shows that an in-depth radio planning is required as these issues can have a great impact in the overall performance of a ISM 2.4 GHz WSN, affecting negatively the potential applications supported by wireless channel. Keywords: judo; wireless sensor networks; ray launching; interference

1. Introduction Nowadays judo is a sport practiced by thousands of people around the world, being one of the most popular martial arts. In addition to the many worldwide competitions that are held under the auspices of the International Judo Federation, judo has been adopted in many countries as part of the physical education programs at different school levels. Thus, there is a huge variety of different training venues where judo is practiced. All these venues usually have some common aspects (such as having a tatami), but in general, each training venue has specific morphological aspects,

Sensors 2016, 16, 1247; doi:10.3390/s16081247

www.mdpi.com/journal/sensors

Sensors 2016, 16, 1247

2 of 25

making them unique complex indoor scenarios in terms of radio propagation due to the size of the venue itself and the presence of different furniture elements (e.g., cupboards, benches, chairs), columns, pads and other elements and equipment that can be found within training gyms, which will have a strong influence in the propagation of the electromagnetic waves as they generate phenomena such as reflection, refraction and diffraction. Therefore, the deployment of WSNs in such environments requires a previous radio planning work, especially considering that electromagnetic interferences created by other wireless sources such as personal portable devices and other wireless systems such as WiFi or other WSNs are likely to happen, making the coexistence in such indoor environments a complex task. Furthermore, in addition to the morphology of the training venues and the potential interferences, the presence of human beings within the scenario is a major issue in terms of radio propagation, especially taking into account that the person density in this kind of scenarios can be high as many judokas can be practicing and exercising at the same time. Even though in the literature many wireless-based applications have been developed for different sport activities such as movement activity monitoring for sprinters [1], arm symmetry investigations for swimmers [2], tracking hip angles for cyclists [3], detection of illegal race walking [4], effort control systems [5] and general monitoring systems based on Ambient Intelligence and Internet of Things (IoT) [6–8], there are very few works related to the deployment of the WSNs within sport venues, and they show WSN deployments within sport venues much larger than judo training venues, such as stadiums [9,10]. When it comes to Wireless Body Area Networks (WBANs) and wearable devices related to sport activities, wearable sensors for training improvement, gait analysis, monitoring sport performance and coaching support have already been developed [11–13], but there are very few that deal with wireless-based applications in martial arts or contact sports [14–16]. Specifically related to judo, which is the sport under analysis in this work, some action recognition works have been published [16,17], but there is no reported work about radio planning analysis nor development of applications for the practice of judo except a radio propagation analysis paper written by the same authors of the present work [18]. Therefore, there is a neediness of in depth radio planning studies for judo training environments, especially taking into account the potential interferences that will be in such environments due to the development of Smart Cities and the IoT, where the radio planning will be a key issue in order to optimize the performance of WSN-based applications. In this work, an in-house developed 3D ray launching algorithm has been employed in order to assess the impact that the presence of human beings has in the radio propagation within a real judo training environment, with the aim of developing potential applications based on wearable wireless devices. For that purpose, in Section 2 the 3D ray launching simulation tool and the description of the scenario under analysis are presented. In Section 3, firstly the validation of the simulation technique is presented. As ZigBee technology has been previously used for other sport monitoring systems [19,20], in this work ZigBee-compliant modules operating at ISM 2.4 GHz band have been used to take measurements within the real judo training scenario with the aim of analyzing the radio propagation as well as for validating the used simulation tool. Then, simulation results for cases with different number of human beings within the scenario are shown, which have been obtained with the inclusion of an in-house developed human body computational model [21]. Additionally, the radio planning analysis is completed with the interference level and Signal to Noise Ratio (SNR) estimations for an emitting WiFi hot spot placed within the scenario. In Section 4 a WiFi and ZigBee-based application for monitoring judokas’ biological constants during training sessions is presented. Finally, in Section 5 the obtained results are discussed. 2. Materials and Methods 2.1. The 3D Ray Launching Method In complex indoor environments like judo training venues, where the morphology of the scenario and the presence of human beings greatly affect the electromagnetic wave propagation, it is a major

Sensors 2016, 16, 1247

3 of 25

issue to conduct a radio propagation analysis before the deploying of a WSN in order to obtain an optimized performance in terms of energy consumption and throughput. Traditionally, empirical methods such as Okumura Hata or COST-231, among others, have been used for that purpose. They provide very rapid results, but due mainly to the multipath propagation phenomenon that happens within complex indoor scenarios, these results are not accurate enough for the scenarios presented in this work, as it is shown later in this work. On the other hand, there are deterministic methods (e.g., Method of Moments and Finite Difference Time Domain), which provide very accurate results. They are based on the resolution of Maxwell’s equations, and the drawback is that they have a high computational cost, i.e., they are highly time-consuming, and therefore, they are used to simulate devices with its immediate surroundings, not a whole environment. As a midpoint, the ray launching and ray tracing methods offer a good trade-off between accuracy and calculation time, providing quite accurate results for a whole environment [22]. In this work, an in-house developed 3D ray launching simulation tool has been used in order to assess the impact that the morphology of the scenario and the presence of human beings have in the radio propagation at 2.4 GHz band within Judo training venues. The presented 3D ray launching algorithm is a deterministic method, based on geometrical optics and geometrical theory of diffraction. The principle of operation of the algorithm lies in the rays that are launched from a specified source in a solid angle (i.e., in 3D directions, both vertical and horizontal) with a predetermined angular separation. Before that, the entire volume of the scenario under analysis is divided into a parameterized mesh with different sizes of cuboids. These cuboids resolution is an input parameter that depends on the dimensions of the scenario. Then, the transmitter antenna is placed at a specific position within the scenario and rays are launched in a specified solid angle. The impact point with an obstacle or with a wall is calculated for each ray, calculating the new angles of the reflected, refracted and diffracted rays. When a ray impinges with a surface of an obstacle, new reflected and refracted rays are generated based on geometrical optics (GO), and when a ray impact with an edge, a new family of diffracted rays are created. These new diffracted rays are based of the geometrical theory of diffraction (GTD) and its extension the uniform theory of diffraction (UTD). While those rays are propagating in the space, all the parameters related with propagation are stored in each cuboid of the defined mesh, for later calculate different results such as received power, power delay profiles, delay spread or signal to noise ratio, among others. For a better comprehension, in Figure 1 this principle of operation is illustrated for a wearable transmitter, as well as the functional diagram of the algorithm. It is important to note that the properties of all the materials of the objects present within the scenario are considered at the frequency of operation (dielectric constant and conductivity). It is also important to mention that the ray launching simulations are mono-frequency, i.e., the results are obtained for a single frequency of operation. The detailed operating mode of the algorithm has been previously published [23], and it has been used and validated in different indoor environments [24–26], including judo environments [18]. For optimum performance of the algorithm, it is highly important to take into account different aspects such as the convergence analysis of the algorithm considering an adequate range for the input parameters. A convergence analysis of the algorithm can be found in [23,27]. These works present the parameter values to be considered in the ray launching approach for optimal performance in terms of accuracy and computational time. The analyzed parameters are the number of reflections, the angular resolution of the launching rays and the number of diffracted rays. It is also worth to mention that for large scenarios the computational time of simulations can be really high depending of the needed accuracy for the results. Because of that, novel hybrid techniques, combining the in-house ray launching approach with different techniques such as neural networks [28], the diffusion equation [29] or collaborative filtering [30] have been also analyzed. These techniques achieve accurate results with considerable lower computational times. The combination of these techniques with the ray launching approach gives an optimized ray launching methodology that is more robust for complex scenarios and it can be used for more applications (i.e., more complex judo training venues with a large number of judokas could be analyzed with such techniques). For better comprehension, Figure 2

Sensors 2016, 16, 1247

4 of 25

Sensors 2016, 16, 1247

4 of 25

Sensors 2016, 16, 1247

4 of 25

shows a scheme of the improvement of the in-house ray launching technique, which have led to the techniques). For better comprehension, Figure 2 shows a scheme of the improvement of the in-house optimal combining different ray approach launching technique, which havetechniques. led to the2 optimal differentoftechniques. techniques). For better comprehension, Figure shows aapproach scheme ofcombining the improvement the in-house ray launching technique, which have led to the optimal approach combining different techniques.

(a) (b) (a) (b) Figure 1. (a) Principle of operation of the in-house 3D ray launching method; and (b) functional Figure 1. (a) Principle of operation of the in-house 3D ray launching method; and (b) functional diagram algorithm. Figure 1. of (a)the Principle of operation of the in-house 3D ray launching method; and (b) functional diagram of the algorithm. diagram of the algorithm.

Figure 2. Performance of the in-house 3D ray launching approach. Figure 2. Performance of the in-house 3D ray launching approach.

Figure 2. Performance of the in-house 3D ray launching approach. 2.2. Scenario Under Analysis 2.2. Scenario Under Analysis TheUnder scenario where the radio propagation analysis has been carried out is a gym located in the 2.2. Scenario Analysis facilities of the Public University Navarre. The dimensions of carried the scenario 18 mlocated (long) in × 8the m The scenario where the radio of propagation analysis has been out isare a gym The scenario whereUniversity the itradio analysis been carried a gym (wide) × of 4.8the m (height) and contains a tatamiThe of 17.5 m × 6 has m,ofwhere judo is practiced. As happens facilities Public ofpropagation Navarre. dimensions the scenario areout 18 mis(long) × 8 located m in the facilities the Public University ofofNavarre. dimensions of the (wide) × 4.8 mof (height) and it contains a tatami 17.5 m × 6 m,The where judo is practiced. As scenario happens are

18 m (long) ˆ 8 m (wide) ˆ 4.8 m (height) and it contains a tatami of 17.5 m ˆ 6 m, where judo is practiced. As happens in many judo training venues, other sports and physical activities are also

Sensors 2016, 16, 1247

5 of 25

Sensors 2016, 16, 1247

5 of 25

practiced, therefore, the scenario contains elements easily in manytherefore, indoor training in many judo training venues, other sportstypical and physical activities arefound also practiced, the scenario contains elements easily many indoor trainingaenvironments, such furniture as a environments, such astypical a punching bags, bigfound foam in balls, doors, columns, mirror and some punching big foam balls, doors, columns, a mirror and some furniture elementsasuch as small elements suchbags, as small lockers and shelves. All these elements make the scenario complex indoor lockers and shelves. All these elements make the scenario a complex indoor environment in terms of environment in terms of radio propagation, as the number of reflections, refractions and diffractions radio propagation, as the number of reflections, refractions and diffractions will be increased, will be increased, making the multipath propagation the main phenomenon within the scenario, making the multipath propagation the main phenomenon within the scenario, much more when much more when persons are present. Figure 3 shows the real scenario and the schematic view of the persons are present. Figure 3 shows the real scenario and the schematic view of the scenario created scenario created the 3D ray launching simulations. for the 3D rayfor launching simulations.

(a)

(b) Figure 3. (a) Real scenario under analysis; (b) Schematic view of the created scenario for 3D ray

Figure 3. (a) Real scenario under analysis; (b) Schematic view of the created scenario for 3D ray launching simulations. launching simulations.

Apart from the real size of the different elements of the scenario, the materials of all the objects within the scenario carefully chosen in order to obtain an electromagnetic behavior as Apart from the realhave sizebeen of the different elements of the scenario, the materials of all the objects closer as possible to the real materials. Table 1 shows the main materials used for the definition of as within the scenario have been carefully chosen in order to obtain an electromagnetic behavior the objects within the scenario, which both their relative permittivity and conductivity at the closer as possible to the real materials. Table 1 shows the main materials used for the definition of the frequency of operation of the wireless system have been obtained from [31]. As the main objective of objects within the scenario, which both their relative permittivity and conductivity at the frequency of this work is to analyze the effect of the presence of persons on the radio wave propagation, the operation of the system havecompleted been obtained from [31]. As main objective of this work is definition of wireless the scenario has been with the inclusion of the an in-house developed human to analyze the effect of the presence of persons on complex the radio wave within propagation, the as definition body computational model [21], which is the most element the scenario the modelof the scenario has been completed with the inclusion of an in-house developed human body computational has been built highly detailed with a broad range of different tissues within it and taking into model [21], which is thesuch mostascomplex element within the scenario as the beenby built highly account body parts skin, different organs, muscles and bones, all model of themhas defined their respective constant (i.e., relative permittivity) and conductivity values. Thus, in addition to as detailed with adielectric broad range of different tissues within it and taking into account body parts such effectmuscles of the presence of people the radio propagation, inclusiondielectric of the human skin, analyzing different the organs, and bones, all ofonthem defined by their the respective constant body model permits theand propagation study of wearable devices, as it can be later inthe thiseffect work. of the (i.e., relative permittivity) conductivity values. Thus, in addition to seen analyzing

presence of people on the radio propagation, the inclusion of the human body model permits the propagation study of wearable devices, as it can be seen later in this work.

Sensors 2016, 16, 1247

6 of 25

Table 1. Material properties defined for the 3D ray launching simulations. Data obtained from [31]. Material

εr

Conductivity (S/m)

Tatami Concrete Aluminum Polypropylene Foam Brick Door Glass Plasterboard

3 25 2.2 3 1.4 4.44 5.84 6.06 2.02

0.2 0.02 37.8 ˆ 106 0.11 0.021 0.11 0.06 10´12 0

Once the scenario for the 3D ray launching has been created, the simulation parameters regarding the launching of rays and the wireless communication such as frequency of operation, antenna gain or cuboids size have to be defined. In Table 2 the main parameters are shown, which have been the same for all the simulations performed in this work. It is worth noting that the parameters have been chosen in order to match those of the real ZigBee-based devices and antennas used for the measurement campaign shown in next section. Table 2. 3D ray launching simulation parameters. Parameter

Value

Frequency Transmitted Power Level Transmitter/Reciever Antenna type Transmitter Antenna Gain Reciver Antenna Gain Launched Ray Resolution (Horizontal and Vertical) Permitted Maximum Reflections Cuboids size

2.41 GHz 10 dBm Monopole 1.5 dBi 7 dBi 1˝ 5 10 ˆ 10 ˆ 10 cm

3. Results In this section, the simulation results for the assessment of the influence that the presence of human beings has on the radio propagation in judo training environments are shown. Firstly, the validation of the 3D ray launching method for its use in the presented scenario has been carried out. For that purpose, a measurement campaign within the real scenario has been made in order to compare these measurement results with those obtained by the 3D ray launching simulations. On one hand, a ZigBee-compliant XBee-Pro module with a 1.5dBi gain monopole antenna mounted on an Arduino board has been used as a transmitter (see Figure 4a). It has been connected to a laptop via USB cable and placed on a standing judoka’s chest, at 1.35 m height. The wireless mote has been configured to operate at 2.41 GHz, which corresponds to the channel 12 of the IEEE 802.15.4 standard. The transmitted power level has been set to 10 dBm and the bitrate to 250 Kbps, with the ‘Serial Interface Data Rate’ set to 125 Kbps (bauds per second). On the other hand, an OAN-1070 monopole 7 dBi gain antenna for ISM 2.4 GHz band operation connected to an Agilent FieldFox N9912A portable spectrum analyzer has been used as receiver. Figure 4 shows both the transmitter mote and the receiving antenna. Before the measurement campaign, a spectrogram has been measured in the center of the scenario under analysis in order to evaluate if any interference could affect the transmission over the chosen wireless channel. It has been measured with the antenna shown in Figure 4b connected to the Agilent FieldFox N9912A portable spectrum analyzer. Figure 5 illustrates the obtained spectrogram. As can be seen, there is not any signal interfering the wireless channel 12, which is centered at 2.41 GHz with a bandwidth of 3 MHz.

Sensors 2016, 16, 1247

7 of 25

Sensors 2016, 16, 1247

7 of 25

Sensors 2016, 16, 1247

7 of 25

(a)

(b)

(a) (b) Figure 4. (a) XBee-Pro module used as transmitter; (b) 2.4–2.5 GHz band monopole antenna used Figure 4. (a) XBee-Pro module used as transmitter; (b) 2.4–2.5 GHz band monopole antenna Figure 4. (a) XBee-Pro module used as transmitter; (b) 2.4–2.5 GHz band monopole antenna usedused as receiver. as receiver. as receiver.

Figure 5. Measured spectrogram within the scenario under analysis for the ZigBee channel 12. Figure 5. Measured spectrogram within the scenario under analysis for the ZigBee channel 12. Figure 5. Measured spectrogram within the scenario under analysis for the ZigBee channel 12. The configuration of the scenario for the radio propagation measurements is represented in The configuration of the scenario forthe thetransmitter radio propagation represented Figure 6, where the standing judoka with on the chestmeasurements can be seen. The is white dashed in The configuration of the scenario for the radio propagation measurements is represented in lines represent linear distance wherethe thetransmitter measurements been can taken. height is 1.35 dashed m, Figure 6, where thethe standing judoka with on have the chest be The seen. The white Figure 6, where the standing judoka with the transmitter on the chest can be seen. The white dashed same height of thedistance transmitter on the chest. The measured power level values areheight depicted in m, lines the represent the linear where the measurements have been taken. The is 1.35 lines represent the linear distance where the measurements have been taken. The height is 1.35 m, Figure 7, where the comparison with the estimations obtained by the 3D ray launching method is in the same height of the transmitter on the chest. The measured power level values are depicted the same height of estimations the transmitter on the chest. The measured power leveland values are depicted in shown. Besides, by traditional empirical models such as COST-231 ITU-R P.1238 are Figure 7, where the comparison with the estimations obtained by the 3D ray launching method is Figure 7, where the comparison the estimations the launching 3D ray launching is included in the graph in order towith compare them with theobtained in-house by 3D ray simulationmethod tool. shown. Besides, estimations by traditional empirical models such as COST-231 and ITU-R P.1238 are shown. Besides, estimations by traditional empirical models such COST-231 and ITU-R P.1238 are As can be clearly seen in Figure 7, the empirical models follow theas tendency of the measurements included in the graph in order to compare them with the in-house 3D ray launching simulation tool. curve,inbut do not fit it properly due tothem the rapid of the received power, which is due tool. included thethey graph in order to compare withvariations the in-house 3D ray launching simulation As can be clearly seen in FigureBesides, 7, the empirical models follow the tendency of the measurements to the multipath propagation. the empirical models do not take into account the effect of the As can be clearly seen in Figure 7, the empirical models follow the tendency of the measurements curve,presence but theyofdo not fit it Thus, properly due to thepower rapidlevels variations of the received power, which isthe due to the judoka. the estimated for x-axis negative values (i.e., behind curve, but they do not fit it properly due to the rapid variations of the received power, which is due the multipath propagation. Besides,Multiwall the empirical models doempirical not take models into account thesince effect of the judoka) obtained by the Cost-231 and ITU-R F.1236 are worse the to the multipath propagation. Besides, the empirical models do not take into account the effect of the losses the human body not takenpower into account Figure negative 7b,c). On the contrary, the 3D the presence ofdue theto judoka. Thus, theare estimated levels (see for x-axis values (i.e., behind

presence of the judoka. Thus, the estimated power levels for x-axis negative values (i.e., behind the judoka) obtained by the Cost-231 Multiwall and ITU-R F.1236 empirical models are worse since the losses due to the human body are not taken into account (see Figure 7b,c). On the contrary, the 3D

Sensors 2016, 16, 1247

8 of 25

judoka) obtained by the Cost-231 Multiwall and ITU-R F.1236 empirical models are worse since the losses due to the human body are not taken into account (see Figure 7b,c). On the contrary, the 3D ray launching algorithm takes into account all the elements within the scenario, including the judoka, 2016, 16, 1247 8 of 25 and its Sensors estimations fit very well the variations observed in measurements. In fact, taken into account Sensors 2016, 16, 1247 8 of 25 the 100 ray measurement points represented in Figure 7, the mean error between measurements launching algorithm takes into account all the elements within the scenario, including the and 3D ray launching estimations is 0.76 dB, with standard deviation of 2.07 These values have been judoka, and its estimations fit very well thea variations observed in measurements. In fact, taken into ray launching algorithm takes into account all the elements within the dB. scenario, including the account thefollowing 100 measurement points represented in Figure 7, thein mean error between measurements obtained by the well fit known formulas: judoka, and its estimations very well the variations observed measurements. In fact, taken into and 3D ray estimations is 0.76 dB, with in a standard of 2.07between dB. These values have account thelaunching 100 measurement points represented Figure the mean error measurements ř7,n deviation been obtained by the following well known x and 3D ray launching estimations is 0.76 dB,formulas: with a standardi“deviation of 2.07 dB. These values have 1 i

Mean Error “ x “

been obtained by the following well known formulas: ∑ n = Mean Error = ̅ d ∑ Mean Error = ̅ = řn

Standard Deviation “ Standard Deviation =

∑ ∑

(1)

(1)

i“1 pxi

( n (

´ xq2 ) )

(1)

(2) (2)

(2)samples where xi is the error between each measurement and =the simulation value and n the number of Standard Deviation where xi is thepoints). error between each measurement and obtained the simulation value and n the number of (100 measurements It is worth noting that the low error is due to the methodology samples measurements points). It is worth noting that the obtained low error is due to the where x i (100 is the error between each measurement and the simulation value and n the number ofis much for obtaining the measurement values. Measuring the received power by a spectrum analyzer methodology obtaining the measurement values. Measuring received power spectrum samples (100 for measurements points). It is worth noting that thethe obtained low errorby is adue to the more accurate than themore RSSIaccurate values provided by values wireless transceivers, which in general give an error analyzer is much than the RSSIvalues. provided by which in methodology for obtaining the measurement Measuring the wireless receivedtransceivers, power by a spectrum from 1 dB to up to 10 dB, depending on the hardware used. As an example, this effect is addressed general an error from 1 dB tothan up tothe 10 dB, on the hardware used. As an example, analyzergive is much more accurate RSSIdepending values provided by wireless transceivers, whichthis in in [32]. effect is give addressed in from [32]. 1 dB to up to 10 dB, depending on the hardware used. As an example, this general an error effect is addressed in [32].

Figure 6. Upper view of the scenario presented in Figure 3 with the position of the transmitter

Figure 6. Upper view of the scenario presented in Figure 3 with the position of the transmitter element element the linear where the measurements have been the taken (white dashed lines). Figure 6.and Upper view distances of the scenario presented in Figure 3 with position of the transmitter and the linear distances where the measurements have been taken (white dashed lines). element and the linear distances where the measurements have been taken (white dashed lines).

(a) (a)

Figure 7. Cont.

Sensors 2016, 16, 1247

9 of 25

Sensors 2016, 16, 1247

9 of 25

(b)

(c)

(d) Figure 7. Linear distance vs. received power level corresponding to the dashed lines depicted in

Figure 7. Linear distance vs. received power level corresponding to the dashed lines depicted in Figure 6: (a) Center; (b) Right side; (c) Left side; (d) Behind columns. Figure 6: (a) Center; (b) Right side; (c) Left side; (d) Behind columns. As the aim of this work is to assess the impact of the presence of persons within this kind of scenario, once the validity of the results obtained by means of the 3D ray launching algorithm for the As the aim of this work is to assess the impact of the presence of persons within this kind of scenario without people (except the judoka with the transmitter) has been made, new measurements scenario, once the validity of the results obtained by means of the 3D ray launching algorithm for the and simulations with the inclusion of an extra human being (2 m in front of the judoka with the scenariotransmitter) without people (except the judoka with transmitter) has been new measurements have been performed in order to the validate the inclusion of themade, in-house developed

and simulations with the inclusion of an extra human being (2 m in front of the judoka with the transmitter) have been performed in order to validate the inclusion of the in-house developed human body model. Figure 8 shows the position of the persons and the new measurement points, represented by white numbers (from 1 to 12). The measurement points are at the height of 1.35 m. Figure 9 shows the comparison between measurements and 3D ray launching simulations. In addition to the results with the inclusion of the person in front of the transmitter, the results without that person are also

Sensors 2016, 16, 1247 Sensors 2016, 16, 1247

10 of 25 10 of 25

human body body model. model. Figure Figure 88 shows shows the the position position of of the the persons persons and and the the new new measurement measurement points, points, human represented by by white numbers (from 1 to 12). The measurement points are at the height of 1.35 m. m. represented Sensors 2016, 16, 1247white numbers (from 1 to 12). The measurement points are at the height of 1.35 10 of 25 Figure 9 shows the comparison between measurements and 3D ray launching simulations. In Figure 9 shows the comparison between measurements and 3D ray launching simulations. In addition to the results with the inclusion of the person in front of the transmitter, the results without addition to the results with the inclusion of the person in front of the transmitter, the results without that person person are also shown. shown. Thus,ofbesides besides the accuracy accuracy of the the obtained estimations, the effectcan of shown. Thus,are besides the accuracy the obtained estimations, theobtained effect of estimations, introducing athe person that also Thus, the of effect of introducing a person can be seen. be seen. introducing a person can be seen.

Figure 8. 8. Upper Upper view view of of the the scenario scenario with with the the new newmeasurement measurement configuration. configuration. The The white white numbers numbers Figure Figure 8. Upper view of the scenario with the new measurement configuration. The white numbers represent the measurement points. represent represent the the measurement measurement points. points.

Figure 9. 9. Measurements Measurements vs. vs. 3D ray ray launching simulation simulation results, both both for the the presence presence of of an an extra extra Figure Figure 9. Measurements vs. 3D 3D ray launching launching simulation results, results, both for for the presence of an extra person and without the person. The measurement points correspond to those shown in Figure 8. person and and without without the the person. person. The The measurement measurement points points correspond correspond to to those those shown shown in in Figure Figure8.8. person

As expected, expected, the the received received power power level level is is lower lower when when the the extra extra person person is is included included in in the the As As expected, the received power level is lower when the extra person is included in the scenario. scenario. It can be also seen that the accuracy of the estimations is also high when the human model scenario. It can be also seen that the accuracy of the estimations is also high when the human model It be alsoinseen that the accuracy of the estimations also high when the human is included is can included the simulated simulated scenario. In order order to gain gainisinsight insight on the the assessment assessment ofmodel the impact impact of the the is included in the scenario. In to on of the of in the simulated scenario. In order to gain insight on the assessment of the impact of the presence of presence of of persons persons within within the the scenario, scenario, further further simulations simulations have have been been performed. performed. For For that that purpose, purpose, presence persons within the scenario, simulations have been For that purpose, XBee-Pro XBee-Pro transmitter has further been placed placed on the the chest of performed. judoka (represented (represented by aa ared red dot in in aa XBee-Pro transmitter has been on chest of aa judoka by dot transmitter has been placed on theachest of a has judoka (represented by asmall red dot in Figure 10), atof1.35 m Figure 10), at 1.35 m height, and receiver been placed on the lockers at height 0.8 m Figure 10), at 1.35 m height, and a receiver has been placed on the small lockers at height of 0.8 m height, and a receiver placed the small alockers at height of 0.8 ma (represented by a red dot (represented by aa red red has dot been in Figure Figure 10),onemulating emulating wireless link between between wearable wireless wireless sensor (represented by dot in 10), a wireless link a wearable sensor in Figure 10), acting emulating a wireless link the between a wearable sensor and a laptop acting as a and laptop as aa sink, sink, receiving information of the thewireless deployed WSN. and aa laptop acting as receiving the information of deployed WSN. sink, receiving the information of the deployed WSN.

Sensors 2016, 16, 1247

11 of 25

Sensors 2016, 16, 1247

11 of 25

(a)

(b)

(c) Figure 10. Upper view of the scenario under analysis with the position of the transmitter and the Figure 10. Upper view of the scenario under analysis with the position of the transmitter and receiver devices (red dots) for the three cases of human being density: (a) Without persons; (b) With the receiver devices (red dots) for the three cases of human being density: (a) Without persons; 10 persons; (c) With 30 persons. (b) With 10 persons; (c) With 30 persons.

Three different cases have been simulated, maintaining both the transmitter and the receiver in Threeplace, different have the been simulated, maintaining the transmitter theare receiver in the same butcases changing person density within theboth scenario. The threeand cases with the the same place, but changing the person density within the scenario. The three cases are with the only presence of the person who has the transmitter, the presence of 10 persons and the presence of only presence the person are whostatic. has the of 10 persons and presence 30 persons. Allofsimulations Thetransmitter, distributionthe of presence the persons throughout thethe scenario has of 30 persons. All simulations are static. The distribution of the persons throughout the been chosen randomly, but taken into account that some of them represent training Judokasscenario (placed has been chosen but taken intooutside accountthe that someFigure of them represent training Judokas on the tatami) andrandomly, others represent people tatami. 10 shows the upper view of the scenario with the three different simulated cases. The simulations with persons have been carried out with the aid of the previously mentioned in-house developed human body model.

Sensors 2016, 16, 1247

12 of 25

(placed on the tatami) and others represent people outside the tatami. Figure 10 shows the upper view of the16,scenario with the three different simulated cases. The simulations with persons have Sensors 2016, 1247 12 been of 25 carried out with the aid of the previously mentioned in-house developed human body model. As itit is is previously previously stated, stated, the the parameters parameters used used for for the the 3D 3DRay RayLaunching Launchingsimulations simulationshave havebeen been As chosen to to match match the the characteristics characteristics of of ZigBee-compliant ZigBee-compliant XBee-Pro XBee-Pro devices devices operating operating at at ISM ISM 2.4 2.4 GHz GHz chosen band. Table Table 22 shows shows them. them. Figure Figure11 11shows showsthe thebidimensional bidimensionalplane planeof ofthe theestimated estimatedreceived receivedpower power band. level at at 0.8 0.8 m m height height (i.e., (i.e., the the height height of the receiver) for the three cases under analysis. Both Both the the level position of of the the transmitter transmitterand andthe thereceiver receiverare arerepresented representedby bywhite whitedots dotswith withthe thetext text‘TX’ ‘TX’and and‘RX’ ‘RX’ position respectively. respectively. Note Note that that the the position position of of the the transmitter transmitter has has been been drawn drawn although althoughits itsreal real position position is is at at height height 1.35 1.35 m. m. The Theshort shortterm termvariation variationof ofthe the received received power power due due to to multipath multipath propagation propagation seen seen in in Figure Figure 77 can can be be also also seen seen throughout throughout the the scenario. scenario. As Asthe themultipath multipathpropagation propagation has has aastrong strong impact impact on on the the received received power powerlevel, level,even evenmore morein inthe thezones zoneswhere wherethe thesurrounding surroundingobstacles obstaclesaffect affectthe theLoS, LoS, two profiles(PDP) (PDP)are arepresented presentedinin Figure case without persons in order to two power delay profiles Figure 1212 forfor thethe case without persons in order to gain gain insight this matter. Specifically, PDP corresponds intransmitter front of the insight in thisinmatter. Specifically, one PDPone corresponds to a point to justainpoint front just of the on transmitter the at Judoka’s chest at m. a distance of 10 The secondtoone corresponds to receiver the pointis the Judoka’son chest a distance of 10 The second onem. corresponds the point where the where is placed Figureseen 10).how It can clearly seen how the first components arrive placedthe (seereceiver Figure 10). It can (see be clearly thebefirst components arrive earlier to the position in earlier thetransmitter position in due fronttoofthe theLoS transmitter due to the LoS and the shorter distance. Besides, less front oftothe and the shorter distance. Besides, less multipath components multipath components lower power levellocation, arrive toasthe location, is expected and with lower power and levelwith arrive to the receiver it isreceiver expected due to as theitsurrounding due to theInsurrounding obstacles. In addition, a to dashed red line is depicted to ashow which obstacles. addition, a dashed red line is depicted show which components have power level components have a powerof level than the sensitivity lower than the sensitivity the lower used XBee-Pro modules. of the used XBee-Pro modules. Regarding the effect effectofofthethe presence of persons, the difference between thewithout case without Regarding the presence of persons, the difference between the case persons persons 30 persons can beseen, clearly as the power receiveddecreases power decreases higher and withand 30 with persons can be clearly as seen, the received at higher atdistances. distances. On the contrary, the difference the case without 10 persons On the contrary, the difference between between the case without persons persons and withand 10 with persons is hard is to hard to note with eye. a naked eye. to Ingain order to gain insight of the ofFigure persons, in note with a naked In order insight in the effectinofthe theeffect presence of presence persons, in 13 the Figure 13 the difference between the received plane for the case and without persons the difference between the received power plane forpower the case without persons the cases withand persons cases with persons is represented. As expected, for the case of 30 persons much is represented. As expected, the difference for thethe casedifference of 30 persons is much bigger than foristhe case bigger than forEven the so, casethe ofdifferences 10 persons. so, of the10differences the case in ofterms 10 persons are of 10 persons. forEven the case persons arefor significant of wireless significant in terms of wireless quality for many points inofthe the decrement of channel quality for many pointschannel in the scenario, as the decrement fewscenario, dB in theasreceived power can few dB in the received power can lead to a failure in the wireless communication due to either a lead to a failure in the wireless communication due to either a received power level lower than the received power lower than the sensibility of the receiver sensibility of thelevel receiver device or an insufficient SNR value. device or an insufficient SNR value.

(a) Figure 11. Cont.

Sensors 2016, 16, 1247 Sensors 2016, 16, 1247

13 of 25 13 of 25

Sensors 2016, 16, 1247

13 of 25

(b) (b)

(c) (c) Figure 11. Estimated received powerlevel leveldistribution distribution plane height 0.8 m m for thethe transmitter on on Figure 11. Estimated Estimated received power planeatatat height transmitter Figure 11. received power level distribution plane height 0.80.8 m forfor the transmitter on the the Judoka’s chest, (a) without morepersons persons in in the the scenario; (b) with 10 10 persons andand (c) with the Judoka’s chest, (a) without more scenario; (b) with persons (c) with Judoka’s chest, (a) without more persons in the scenario; (b) with 10 persons and (c) with 30 persons. 30 persons. 30 persons.

(a)

(b)

Figure 12. Power delay profiles at two different positions: (a) In front of the transmitter on the

Figure 12. Power delay on the judoka’s (a) profiles at two different positions: (a) In front of the transmitter (b) judoka’s chest at a distance of 10 m; (b) At the receiver location. chest at a distance of 10 m; (b) At the receiver location. Figure 12. Power delay profiles at two different positions: (a) In front of the transmitter on the judoka’s chest at a distance of 10 m; (b) At the receiver location.

Sensors 2016, 16, 1247 Sensors 2016, 16, 1247

14 of 25 14 of 25

(a)

(b) Figure 13. Received power level difference for the bidimensional plane at 0.8 m height between the Figure 13. Received power level difference for the bidimensional plane at 0.8 m height between the results without persons and (a) with 10 persons; (b) with 30 persons. results without persons and (a) with 10 persons; (b) with 30 persons.

Once the effect of the presence of persons within the scenario under analysis in terms of Once the effect of has the presence of persons within scenario in terms of received power level been performed, how thesethe results canunder affect analysis the deployment of received wireless power level has howand these resultsconsumption can affect the is deployment wireless transceivers transceivers in been termsperformed, of data rate energy presented.ofFirstly, the estimated in terms ofpower data rate andgive energy is presented. Firstly, the estimated received power values received values the consumption information needed in order to know if a specific wireless transceiver give information neededminimum in order tosignal know if a specific wireless transceivercommunication will receive the with required will the receive the required power to have a successful the minimum signal power to have a successful communication with the transmitter. This minimum transmitter. This minimum signal power is given by the sensitivity of the transceivers. In this case, signal power is given byhave the sensitivity ofof the transceivers. In this case, the by XBee-Pro modules havefor a the XBee-Pro modules a sensitivity −100 dBm, which is surpassed the received power sensitivity ´100 dBm, which is surpassed by scenario. the received almost all the almost all of the points throughout the whole Butpower this isfor not enough to points have athroughout successful the whole scenario. But this is not enough to have a successful wireless communication between the wireless communication between the transmitter and the potential receiver, as electromagnetic transmitter the potential receiver, as electromagnetic interference likelycontext to be present such a interferenceand is likely to be present in such a scenario, even more in aisfuture awareinscenario scenario, even a future aware scenarioIn framed IoT the andimpact Smart City framed by themore IoT in and Smartcontext City environments. orderby to the show thatenvironments. interferences In order to show the impact that interferences could have on the wireless communication within judo could have on the wireless communication within judo training environments, as an example, four training environments, as an example, four WiFi access points have been placed in the scenario, fixed WiFi access points have been placed in the scenario, fixed to the ceiling (height of 3.9 m), emitting to ceiling of 3.9 m),ofemitting 20ofdBm at the same frequency of operation of power the ZigBee 20 the dBm at the(height same frequency operation the ZigBee motes. The ZigBee transmitted level motes. The ZigBee transmitted power level has been set to 10 dBm. The location of the WiFi has been set to 10 dBm. The location of the WiFi access points as well as the estimated WiFi access power points as wellatas the estimated distribution the heightinofFigure 0.8 m 14, for for each access point is distribution the height of 0.8WiFi m forpower each access point isatpresented the case without presented in Figure 14, for the case without persons. In order to assess if the ZigBee communication persons. In order to assess if the ZigBee communication can be successfully achieved when those can successfully achieved when those WiFi access points transmitting, the relation WiFibeaccess points are transmitting, the relation between the are received ZigBee signal powerbetween and the the received produced ZigBee signal power thepoints interference produced by the access points been interference by the WiFiand access has been calculated, i.e.,WiFi the SNR. Note thathas both the calculated, i.e., the SNR. Note that both the WiFi access points and the ZigBee motes usually transmit WiFi access points and the ZigBee motes usually transmit traffic burst, not continuously. Therefore, traffic burst, not continuously. Therefore, the interference between when those two happens the interference between those two wireless systems happens bothwireless systemssystems transmit at the same time, i.e., when collision of both signals happen. The SNR has been calculated by the following well known formula:

Sensors 2016, 16, 1247

15 of 25

when both systems transmit at the same time, i.e., when collision of both signals happen. The SNR has been calculated by the following well known formula: C “ BW ˆ log2 p1 ` S{Nq Sensors 2016, 16, 1247

(3) 15 of 25

where C is the channel capacity in bps (250 Kbps fixed for ZigBee), BW is the communication system bandwidth in Hz (3 MHz for ZigBee) and S and N are the power levels of the received signal and C = BW × log2(1 + S/N) (3) noise respectively (in Watts). The required minimum SNR for a successful ZigBee communication wheredB. C isTable the channel capacity in bps (250 results Kbps fixed for ZigBee), is the communication system is ´12.26 3 shows the simulation of the receivedBW ZigBee signal power level as well bandwidth in Hz (3 MHz for ZigBee) and S and N are the power levels of the received signal and as WiFi interference levels at the receiver location (Rx), and in Figure 15 the estimated SNR values noise respectively (in Watts). The required minimum SNR for a successful ZigBee communication is at receiver position for the different WiFi access point positions are depicted. The dashed red line −12.26 dB. Table 3 shows the simulation results of the received ZigBee signal power level as well as represents the previously calculated minimum SNR value of ´12.26 dB. As can be seen, the potential WiFi interference levels at the receiver location (Rx), and in Figure 15 the estimated SNR values at positions of position the wireless transceivers, both the point motespositions of our network andThe the dashed interfering devices, receiver for the different WiFi access are depicted. red line haverepresents a great impact on thecalculated performance, which theofsame will strongly on the the previously minimum SNR at value −12.26time dB. As can depend be seen, the potential morphology positions of of the the scenario. wireless transceivers, both the motes of our network and the interfering devices, have a great impact on the performance, which at the same time will depend strongly on the Table 3. Simulation results at receiver point (Rx) for ZigBee transmitting 10 dBm and WiFi morphology of the scenario. transmitting 20 dBm. Table 3. Simulation results at receiver point (Rx) for ZigBee transmitting 10 dBm and WiFi transmitting 20 dBm. WiFi Location ZigBee Signal Power (dBm) WiFi Interference Power (dBm)

WiFi WiFi Location 1 WiFiWiFi 2 1 WiFiWiFi 3 2 WiFiWiFi 4 3 WiFi 4

ZigBee Signal ´64.12 Power (dBm) ´64.12 −64.12 ´64.12 −64.12 ´64.12 −64.12 −64.12

WiFi Interference Power (dBm) ´51.16 −51.16 ´71.11 −71.11 ´68.99 −68.99 ´48.91 −48.91

Figure 14. Cont.

Sensors 2016, 16, 1247 Sensors 2016, 16, 1247 Sensors 2016, 16, 1247

16 of 25 16 of 25 16 of 25

Figure 14. WiFi power distributionfor forthe thebidimensional bidimensional plane 0.80.8 mm height for for the the casecase without Figure 14. WiFi power distribution planeatat height without persons. The red dots represent the WiFi access point locations. persons. The red dots represent the WiFi access point locations. Figure 14. WiFi power distribution for the bidimensional plane at 0.8 m height for the case without

persons. The red dots represent the WiFi access point locations.

Figure 15. Estimated SNR values at receiver position for different WiFi access point positions, for different ZigBee transmitted power levels.

In order to gain insight in how the presence of human beings affects the performance of a Figure 15. Estimated SNR values at at receiver position for for different WiFi access point positions, for Figure Estimated SNR values receiver position different point positions, ZigBee 15. communication in terms of SNR, the WiFi 4 position has been WiFi takenaccess as an example and in different ZigBee transmitted power levels. for different transmitted levels.position (Rx) for the three cases without persons, with Figure 16 theZigBee SNR calculated at power the receiver 10 persons and with 30 persons is presented. The x-axis indicates different transmission power levels In order to gain insight in judoka’s how thechest. presence of human beings affects the performance of a fororder the ZigBee of in thehow that although the European regulationsof allow In to gainsensor insight the presenceNote of human beings affects the performance a ZigBee ZigBee communication terms of SNR, the WiFi transmitting 4 position has been taken asanalysis an example in transmitting upterms to 10in dBm, thethe inclusion higher power in the due and to16 the communication in of SNR, WiFi 4ofposition has been taken aslevels an example and in is Figure Figurethe16limitations the SNR are calculated at the receiver position (Rx) for the three cases without persons, with different in some other parts of the world and there are commercial devices which SNR calculated at the receiver position (Rx) for the three cases without persons, with 10 persons and 10 persons and with 30 persons is presented. The x-axis indicates different transmission power levels can transmit higher power level (e.g., the XBee-Pro modules used in this work, which transmit up to with 30 persons is presented. The x-axis indicates different transmission power levels for the ZigBee 18 ZigBee dBm). Table 4 shows thejudoka’s simulation results of the received signalthe power levels as well as WiFi for the sensor of the chest. Note that although European regulations allow

sensor of the judoka’s chest. Note that although the European regulations allow transmitting up transmitting up to 10 dBm, the inclusion of higher transmitting power levels in the analysis is due to to 10 dBm, the inclusion of higher transmitting power levels in the analysis is due to the limitations the limitations are different in some other parts of the world and there are commercial devices which are different in some other parts of the world and there are commercial devices which can transmit can transmit higher power level (e.g., the XBee-Pro modules used in this work, which transmit up to higher power level (e.g., the XBee-Pro modules used in this work, which transmit up to 18 dBm). 18 dBm). Table 4 shows the simulation results of the received signal power levels as well as WiFi

Sensors 2016, 16, 1247

17 of 25

Sensors 2016, 16, 1247

17 of 25

Table 4 shows the simulation results of the received signal power levels as well as WiFi interference levels at the levels receiver location (Rx)location when ZigBee transmits 10transmits dBm. As10 can be seen, forbe the WiFifor 4 interference at the receiver (Rx) when ZigBee dBm. As can seen, position, receivedthe WiFi signal WiFi is notsignal affected by the inclusion of inclusion persons, which is mainly due the WiFi the 4 position, received is not affected by the of persons, which is to the shorter distance between WiFi access point andpoint the receiver. But the received ZigBee mainly due to the shorter distancethe between the WiFi access and the receiver. But the received signal affected significantly, which iswhich expected as in theas case under analysis the included persons ZigBeeissignal is affected significantly, is expected in the case under analysis the included are in theare path ZigBee transmitter and the receiver. to Figure the difference persons in between the path the between the ZigBee transmitter and the Back receiver. Back 16, to Figure 16, the between cases without persons andpersons including persons is due mainlyistodue the mainly lower ZigBee differencethe between the cases without and including persons to the signal lower power at thereceived Rx point,atwhich to awhich lowerlead SNRtovalue. As SNR expected, increases ZigBee received signal power the Rxlead point, a lower value.the AsSNR expected, the when the transmitting power levels arepower higher.levels But itare is worth noting 250noting Kbps data SNR increases when the transmitting higher. But itthat is the worth that rate the is not achievable transmitting 10 dBmtransmitting or less when10 collision thecollision WiFi andbetween ZigBee happens. 250 Kbps data rate is not achievable dBm orbetween less when the WiFi Increasing transmitting power above 10 power dBm will avoid that at the of and ZigBeethe happens. Increasing thelevel transmitting level above 10problem, dBm willbut avoid thatexpense problem, abut higher consumption the transmitter, which is likely to be powered batteries. Instead, at theenergy expense of a higher of energy consumption of the transmitter, which is by likely to be powered the proposed Instead, method based on the 3Dmethod ray launching algorithm in finding an optimized WSN by batteries. the proposed based on the 3D can rayaid launching algorithm can aid in deployment, in order WSN to obtain optimized placement in terms of achievable datainrate as finding an optimized deployment, intransceiver order to obtain optimized transceiver placement terms well as energydata consumption. of achievable rate as well as energy consumption.

Figure 16. 16. Estimated EstimatedSNR SNRvalues valuesfor for different configurations of the Tx-Rx wireless link when the Figure different configurations of the Tx-Rx wireless link when the WiFi WiFi 4 access point is interfering. 4 access point is interfering. Table 4.4. Simulation WiFi Table Simulation results results at at receiver receiver point point (Rx) (Rx) for for ZigBee ZigBee transmitting transmitting 10 10 dBm dBm and and WiFi transmitting 20 20 dBm. dBm. transmitting Human Density Case Human DensityPersons Case Without Without 10Persons Persons 10 Persons 30 Persons 30 Persons

ZigBee Signal Power (dBm) WiFi Interference Power (dBm) ZigBee Signal WiFi Interference Power (dBm) −64.12 Power (dBm) −48.91 ´64.12 −77.85 ´77.85 −86.08 ´86.08

−48.91 ´48.91 −49.02 ´48.91 ´49.02

Other potential communication systems could be employed in order to provide the required connectivity, such as communication WLAN or mobilesystems communication In in general coverage/capacity Other potential could be systems. employed orderterms, to provide the required relations for low transmission in the same range as 250 Kbps) would given the fact connectivity, suchbit asrate WLAN or mobile(i.e., communication systems. In general terms,hold, coverage/capacity that transmission power can be higher(i.e., in both cases. new coverage would relations for low bit rate transmission in the sameEventually, range as 250 Kbps) wouldanalysis hold, given thehave fact to betransmission carried out to analyze parameters holdnew as acoverage function analysis of increased bithave rate that power can ifbequality higher of in service both cases. Eventually, would demand, a situation that could happenof in service servicesparameters such as realhold timeas video are required. This isbit a topic to be carried out to analyze if quality a function of increased rate for futureawork, as athat function required end-usersuch application demands. demand, situation couldofhappen in services as real time video are required. This is a topic for future work, as a function of required end-user application demands. 4. Judoka Monitoring Application 4. Judoka Monitoring Application In [18] we developed an application aimed at monitoring the practice of judo focusing on In [18]the we tasks developed application at monitoring the practice judo focusing on facilitating of judoanarbitration foraimed the referee and to the two cornerofjudges, helping the facilitating the tasks of judo arbitration for the referee and to the two corner judges, helping the audience at competitions to understand the fighting and the scores given by the referees, and allowing the refinement of judo waza. Following this line of work, we have developed an application that monitors certain biological parameters of the judokas and oversees their efforts at training and competition, in order to improve the quality of training, while ensuring that there is not

Sensors 2016, 16, 1247

18 of 25

audience at competitions to understand the fighting and the scores given by the referees, and allowing the refinement of judo waza. Following this line of work, we have developed an application that monitors certain biological parameters of the judokas and oversees their efforts at training and competition, in order to improve the quality of training, while ensuring that there is not conducted un-recommended overexertion. The developed application uses commercial off-the-shelf (COTS) products. It has been designed to use standard manufactured products rather than customized, or bespoke, products. We have focused on minimizing the cost of the hardware equipment, to grant an easy integration of the different commercial equipment, and on developing a software that takes full advantage of the system with lowest energy consumption. The system integrates standard and well known tools. Judo requires, as do other sports, the combination of a good physical preparation of both anaerobic and aerobic type, since a high resistance to withstand the duration of a bout as well as explosive and high-speed actions is required. When properly performed, Judo training can provide significant functional benefits and improvement in overall health and technique. We can distinguish three different phases needed to produce the big performance desired by a judoka when planning his/her training: the preparatory, competitive and transit periods. During the preparatory period, strength, speed, endurance, agility, flexibility, mobility and technical issues are addressed. The competitive period is devoted to increase the intensity of the activity, while the transit period is devoted to the functional regeneration of the judoka. In order to facilitate the collection of information that enables better planning and executing of training we have developed the application described below. The technical novelty of the proposal is the simultaneous monitoring of multiple judokas. When workouts are collective, the difficulty of monitoring the training conditions of all judokas makes it impractical for just a person, or for a reduced number of them, to supervise the activity of multiple judokas at a same time. For this reason it is necessary to propose a system architecture that grants those objectives are met. A noninvasive method has been followed for monitoring the judoka’s oxygen saturation (SO2 ) by reading his/her peripheral oxygen saturation (SpO2 ), while his/her heart rate is also measured. Figure 17 shows the devices involved in the sensing process, while Figure 18 describes the software architecture of the system, which consists of three layers: sensor, service and application layers. The device in charge of data sensing is a Waspmote node [33], which aggregates the data collected by a SpO2 sensor and a heart rate monitor and summits it to a web service in charge of permanent data storing. The sensor layer is located at the Waspmote node and includes a communication module, which allows data collection from sensor devices by means of a Bluetooth connection and WiFi communication with the service layer. The node provides a certain storage capacity, which is provided by an SD card. The node can be programmed by a small configuration module (Config) to interrogate the available sensors and acquire the corresponding information. This information is stored in the SD card through the local data storage module, and analyzed by a reasoner. The reasoner is responsible of discerning whether the values sensed suggest that a judoka is reaching a worrying level, and if so, it urgently notifies this fact to the service layer. The setup of the nodes establishes the refresh rate of the sensors. Internally, the node acquires information continuously from sensors, since it performs a continuous loop data refresh (data garbage). Data is always stored locally into the SD card and transmitted to the storage web service according to the network availability. If the reasoner does not observe any significant abnormality after evaluating each sensed data, it requests the delayed persistent storage of the data to the storage web service. The delayed transmission allows a better use of the available bandwidth, minimizing the transmission time in order to avoid collisions between bursts of messages. The node aggregates the information available and looks forward periods of low network activity to send the information to the correspondent web service. If the reasoner does not observe any significant abnormality after evaluating the data obtained, it requests the persistent storage of the data. If the reasoner appreciates any anomaly, it urgently informs this fact to the web notification service located at the service layer in order to take the required actions such as stopping

Sensors 2016, 16, 1247

19 of 25

the training session. Periodically, where this refresh rate can be modified by users when needed (Config module), the node sends the data aggregated during the last sensing period to the data collection service located at the sensor layer. This service stores the data collected into a relational database, in our case, we use a MySQL database server. All the modules of the sensor layer are implemented as pieces of code embedded in the cyclic algorithm executed by the node over its operating Sensors 2016,system. 16, 1247 19 of 25 Sensors 2016, 16, 1247

19 of 25

Figure 17. Hardware devices involved in the monitoring process. Figure 17. 17. Hardware Hardware devices devices involved involved in in the the monitoring monitoring proces process. Figure s.

Figure 18. Software architecture of the monitoring system. Figure of the the monitoring monitoring system. system. Figure 18. 18. Software Software architecture architecture of

The service layer implements all its functionalities by means of web services. The three services The service layer implements all its functionalities by means of web services.The The three services provided (programming, data collection and notification) are of web The service layer implements all its functionalities by means webservices. services. The programming three services provided (programming, data collection and notification) are web services. The programming services allow an easy and quick configuration of the sensors andservices. their associated parameters. Thus, provided (programming, data collection and notification) are web The programming services services allow an easy and quickaconfiguration of the sensors and their associated parameters. Thus, the user can determine through web service what sensors correspond to what judoka, how often to allow an easy and quick configuration of the sensors and their associated parameters. Thus, the user the user can determine through a web service what sensors correspond to what judoka, how often to monitor the acquired which alert thresholds should the reasoner take into account, and many can determine throughdata, a web service what sensors correspond to what judoka, how often to monitor monitor the acquired data, is which alert executing thresholdsthat should the reasoner take into account, andservice many others. Note that the node cyclically algorithm forever. The data collection the acquired data, which alert thresholds should the reasoner take into account, and many others. others. Note that thedata nodeinto is cyclically executing that algorithm collection service stores the received a database and acknowledges theforever. storageThe to data the sensor layer. This stores the received data into a database and acknowledges the storage to the sensor layer. This ensures that no information is lost, since the storage module periodically requests to the storage ensures thatstorage no information lost, since storage module periodically to the service the of all data is pending. If a the failure occurs when invoking the requests web service, thestorage sensor service the storage of all data pending. If a failure when web service, theand sensor will re-request the storage in the next iteration andoccurs will not giveinvoking up until the it reaches its target the will re-request the storage in the next iteration and will not give up until it reaches its target and the proper storage of information into the database is acknowledged. The notification service is in proper storage of information the(reasoner) database to is the acknowledged. The notification is in charge of alert notification from into sensors application layer, although theservice information

Sensors 2016, 16, 1247

20 of 25

Note that the node is cyclically executing that algorithm forever. The data collection service stores the received data into a database and acknowledges the storage to the sensor layer. This ensures that no information is lost, since the storage module periodically requests to the storage service the storage of all data pending. If a failure occurs when invoking the web service, the sensor will re-request the storage in the next iteration and will not give up until it reaches its target and the proper storage of information into the database is acknowledged. The notification service is in charge of alert notification from sensors (reasoner) to the application layer, although the information is also transferred to the storage service for its persistent storage. Those services, as well as the MySQL 5.6 and Tomcat 6 servers, are located at 1247 a laptop and communicate through a WiFi LAN network. The location of the servers Sensors 2016, 16, 20 of 25 Sensors 2016, 16, of 25 in a laptop is1247 due to mobility and portability reasons. We have tried to make the system the20most portable In Figure 19 the well-known followed by to thediscover three services described followedpossible. by the three services described abovearchitecture and the elements used (UDDI), publish followed by the three services described above and the elements used to discover (UDDI), publish above and the elements used to discover (UDDI), publish (WSDL) and invoke (SOAP) web services (WSDL) and invoke (SOAP) web services are depicted. Figure 20 shows the exceptions that can be (WSDL) and invoke (SOAP) web are depicted. Figure 20by shows the exceptions that can be are depicted. Figure 20 shows theservices exceptions can beA thrown the reasoner to the notification thrown by the reasoner to the notificationthat service. Periodical Exception just implies a new thrown by the reasoner to the notification service. A Periodical Exception just implies a new service. Periodical Exception justException implies aimplies new storage iteration, while anas Issue implies a storage A iteration, while an Issue a certain problem such the Exception notification of any storageproblem iteration,such while an Issue Exception implies a certain problem such as the the notification of any certain as the notification of any previous communication problem, storage request previous communication problem, the storage request of pending data from previous failed storage previous communication problem, the storage request of pending data from previous failed storage of pending cycles, etc. data from previous failed storage cycles, etc. cycles, etc.

Figure 19. 19. Web Web service architecture. architecture. Figure 19. Figure Web service service architecture.

Figure 20. Possible issue notifications by the reasoner. Figure Figure 20. 20. Possible Possible issue issue notifications notifications by by the the reasoner. reasoner.

Finally, the Alert Exception implies the urgent notification to the application layer of a serious Finally, the Alert Exception implies the urgent notification to the application layer of a serious incident suchthe as Alert the total or partial loss of connectivity, or something important, such any of Finally, Exception implies the urgent notification to themore application layer of aasserious incident such as the total or partial loss of connectivity, or something more important, such as any of the judokas exceeded any ofloss theofrisk thresholds established. In such all cases, incident such has as the total or partial connectivity, or previously something more important, as anyand of the judokas has exceeded any of the risk thresholds previously established. In all cases, and according to the priority levels defined (alert, issue and period) the service layer notifies these the judokas has exceeded any of the risk thresholds previously established. In all cases, and according according to the priority levels defined (alert, issue and period) the service layer notifies these exceptions to levels the application layer. Much of the the potential is minimized due to the to the priority defined (alert, issue and period) servicetraffic layer notifies these exceptions to exceptions to the application layer. Much of the potential traffic is minimized due to the pre-processing performed sensors. The traffic carried is reduced, which prevents flooding and the application layer. Muchon of the potential is traffic minimized due to the pre-processing performed pre-processing performed on sensors. The carried traffic is reduced, which prevents flooding and minimizes collision a collective training, which several channel judokas collision compete on sensors. channel The carried trafficprobability. is reduced, In which prevents floodinginand minimizes minimizes channel collision probability. In a collective training, in which several judokas compete simultaneously, multiple bursts of traffic between the sensors and the rest of the system must be probability. In a collective training, in which several judokas compete simultaneously, multiple bursts simultaneously, multiple bursts of traffic between the sensors and the rest of the system must be considered. In addition, since a judoka against we frequently observe both of traffic between the sensors and the rest competes of the system mustanother, be considered. In addition, sincethat a judoka considered. In addition, since a judoka competes against another, we frequently observe that both sensors try to communicate at the same time with the system, as both competitors simultaneously competes against another, we frequently observe that both sensors try to communicate at the same sensors try to communicate at the same time with the system, as both competitors simultaneously perform high-intensity efforts. For such reason, and to minimize the workload of the monitor, the perform high-intensity efforts. For such reason, and to minimize the workload of the monitor, the filtering task performed by the reasoner is so relevant. filtering task performed by the reasoner is so relevant. Finally, the application layer, which is implemented following the Model, View, Controller Finally, the application layer, which is implemented following the Model, View, Controller (MVC) pattern, mainly concerns four issues: frontend, reporting, monitoring and configuring. (MVC) pattern, mainly concerns four issues: frontend, reporting, monitoring and configuring. Following that scheme, an Android-based app has been developed, which is depicted in Figure 21.

Sensors 2016, 16, 1247

21 of 25

time with the system, as both competitors simultaneously perform high-intensity efforts. For such reason, and to minimize the workload of the monitor, the filtering task performed by the reasoner is so relevant. Finally, the application layer, which is implemented following the Model, View, Controller (MVC) pattern, mainly concerns four issues: frontend, reporting, monitoring and configuring. Following that scheme, an2016, Android-based app has been developed, which is depicted in Figure 21. Sensors 16, 1247 21 of 25

Figure 21. View of theofpresented monitoring appapp running in aintablet. TheThe single (on(on thethe tablet’s screen) Figure 21. View the presented monitoring running a tablet. single tablet’s screen) and dualmonitoring (zoomed) monitoring of aare randori are represented. and dual (zoomed) of a randori represented.

The aim of the application is to monitor different biological parameters of judokas during their

The aim of the application is to monitor different biological parameters of judokas during their training exercises, such as randori (i.e., combat). The application allows the monitoring of a single training exercises, as randori (i.e., combat). The application allows monitoring a single judoka or a set such of them. It also allows the simultaneous comparison of twothe judokas, usuallyof those judoka or a set ofare them. It also allows simultaneous comparison of two usually judokas who fighting against eachthe other. The user can select the judoka to judokas, be monitored fromthose judokas whoallare fighting against each The user can select to application be monitored among those available. While thisother. monitoring is performed in the realjudoka time, the alsofrom allows re-displaying previous already stored the database. Thetime, frontend is in charge among all those available. While data this monitoring is in performed in real the module application also allows of user login, data connection, user registration, the database. election of the to be monitored and the of re-displaying previous data already stored in the Thejudoka/s frontend module is in charge customization of the graphic user interface. The monitoring module, depicted in Figure 21, is in the user login, data connection, user registration, the election of the judoka/s to be monitored and charge of data comparison, alert notification to users and the real-time presentation of the data customization of the graphic user interface. The monitoring module, depicted in Figure 21, is in charge collected by sensors. The configuration module allows the setup of sensors, the enabling/disabling of of data comparison, alert notification to users and the real-time presentation of the data collected by sensors, the registration of new devices in the system, the publishing of new services, etc. The reporting module makes use of the Jasper Reports tool to provide powerful and useful reports aimed at improving the training methodology of judokas. The reporting module makes use of the Edrawsovt 2015 v7.9 tool to provide powerful and useful reports aimed at improving the training

Sensors 2016, 16, 1247

22 of 25

sensors. The configuration module allows the setup of sensors, the enabling/disabling of sensors, the registration of new devices in the system, the publishing of new services, etc. The reporting module makes use of the Jasper Reports tool to provide powerful and useful reports aimed at improving the training methodology of judokas. The reporting module makes use of the Edrawsovt 2015 v7.9 tool to provide powerful judokas. Sensors 2016, 16, 1247 and useful reports aimed at improving the training methodology of 22 of 25 Reports provided aid to supervise and improve the training process. The user can select the type methodology of judokas. Reports provided aid to supervise and improve training process. The and number of indicators and graphs, allowing the customization of thethe reports provided. All the user can select the type and number of indicators and graphs, allowing the customization of the information gathered along the time is achieved into a database, so the system allows a wide number of reports provided. All the information gathered along the time is achieved into a database, so the individual and collective comparatives studies according to the weight divisions, the gender, category system allows a wide number of individual and collective comparatives studies according to the (junior, senior . . . ), etc. weight divisions, the gender, category (junior, senior…), etc.

Figure 22. Reporting examples: individual report (top) and collective report by category (bottom).

Figure 22. Reporting examples: individual report (top) and collective report by category (bottom).

Figure 22 illustrates some of the reports provided. The top figure shows the evolution of the training process of a certain judoka while the bottom part shows the evolution comparison among

Sensors 2016, 16, 1247

23 of 25

Figure 22 illustrates some of the reports provided. The top figure shows the evolution of the training process of a certain judoka while the bottom part shows the evolution comparison among the four judokas belonging to the category of less than 81 Kg. As it can be observed in Figure 22 (bottom), judoka #4 has been injured for more than three months. One can note that the highest degree of training/competitiveness corresponds to the celebration of the European Judo Championship 2016 celebrated in Kazan (21 April, Russia). Right now, all the reports are generated on demand and stored into a private repository, while our future works focus on the use of a business intelligence tool as Pentaho to design and provide interactive dashboards. 5. Conclusions In this work, the influence of the presence of human beings on the performance of WSNs at ISM 2.4 GHz band in judo training venues is analyzed by means of an in-house developed 3D ray launching algorithm, with the aid of an in-house developed human body computational model. The obtained results show the typical short term variations of the received power level due to the multipath propagation, which is usually the strongest propagation phenomenon in this kind of indoor environments. The comparison between simulation results and measurements show that the in-house developed 3D ray launching method is an accurate tool in order to obtain received power estimations within judo training environments, as the estimations fit the short term variations, in contrast to traditional empirical methods, which only give the tendency of the values. The presented results show that in addition to the placement of the wireless transceivers (transmitters as well as receivers), which has a great impact in the power distribution throughout the scenario, the presence of persons is a key issue that has to be taken into account in order to obtain an optimized deployment of a ZigBee WSN in judo training environments. In fact, the high density of persons that can be found in this kind of scenarios can be a determining factor in the performance of the deployed ZigBee-based WSN in terms of achievable data rate and energy consumption. The presented method can aid in obtaining the optimal wireless network deployment, configuration and performance, making the use of WSNs attractive for the adoption of applications within judo training venues, such as the biological constant monitoring application presented in this work. This method can be transferred for similar assessments of any judo training venue as well as other sport venues with similar morphological characteristics, where besides judo, other different sports could be practiced, such as other martial arts and diverse physical activities (e.g., yoga, gymnastics for seniors, any kind of dance or aerobics, just to name a few). As future work, the effect of the dynamic presence of human beings within this kind of scenarios as well as the analysis of other wireless technologies and operating frequencies will be interesting issues to analyze. Regarding the developed application, the software architecture proposed allows the implementation of new applications using other technologies or platforms (HTM5, IOS ...) with the least effort. Future works concern the integration of two tools as Weka and Pentaho in order to better define the rules of the reasoner and provide precise dashboards, respectively. Acknowledgments: The authors wish to acknowledge the financial support of project TEC2013-45585-C2-1-R, funded by the Spanish Ministry of Economy and Competiveness. The authors wish to thank the collaboration of Judo Klub Erice and the Public University of Navarre, for providing their facilities. Author Contributions: Peio Lopez-Iturri, Erik Aguirre, Leyre Azpilicueta and Francisco Falcone have designed and performed the experiments and simulations for the wireless system analysis; Jose Javier Astrain and Jesús Villadangos have developed the Judoka monitoring application. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations The following abbreviations are used in this manuscript: WSN SNR ISM IoT

Wireless Sensor Network Signal to Noise Ratio Industrial, Scientific and Medical Internet of Things

Sensors 2016, 16, 1247

24 of 25

References 1.

2. 3.

4. 5.

6.

7.

8.

9.

10.

11.

12.

13. 14. 15.

16.

17.

Kuznietsov, A.; Neubauer, D. A wireless framework for movement activity monitoring of sprinters. In Proceedings of the 9th International Multi-Conference on Systems, Signals and Devices, Chemnitz, Germany, 20–23 March 2012; pp. 1–6. Stamm, A.; James, D.A.; Hagem, R.M.; Thiel, D.V. Investigating arm symmetry in swimming using inertial sensors. In Proceedings of the IEEE Sensors, Taipei, Taiwan, 28–31 October 2012; pp. 1–4. Cockcroft, J.; Muller, J.H.; Scheffer, C. A Novel Complimentary Filter for Tracking Hip Angles During Cycling Using Wireless Inertial Sensors and Dynamic Acceleration Estimation. IEEE Sens. J. 2014, 14, 2864–2871. [CrossRef] Lee, J.B.; Mellifont, R.B.; Burkett, B.J.; James, D.A. Detection of Illegal Race Walking: A Tool to Assist Coaching and Judging. Sensors 2013, 13, 16065–16074. [CrossRef] [PubMed] Vales-Alonso, J.; Lopez-Matencio, P.; Veiga-Gontan, J.; Banos-Guirao, P.; Alcaraz, J.J. An effort control system for training elite team-sport athletes. In Proceedings of the 6th International Conference on Human System Interaction, Sopot, Poland, 6–8 June 2013; pp. 279–286. Rowlands, D.D.; McNab, T.; Laakso, L.; James, D.A. Cloud based activity monitoring system for health and sport. In Proceedings of the International Joint Conference on Neural Networks, Brisbane, Australia, 10–15 June 2012; pp. 1–5. Vales-Alonso, J.; Lopez-Matencio, P.; Gonzalez-Castaño, F.J.; Navarro-Hellin, H.; Baños-Guirao, P.J.; Perez-Martinez, F.J.; Martinez-Alvarez, R.P.; Gonzalez-Jimenez, D.; Gil-Castiñeira, F.; Duro-Fernandez, R. Ambient Intelligence Systems for Personalized Sport Training. Sensors 2010, 10, 2359–2385. [CrossRef] [PubMed] Rodríguez-Molina, J.; Martínez, J.-F.; Castillejo, P.; López, L. Combining Wireless Sensor Networks and Semantic Middleware for an Internet of Things-Based Sportsman/Woman Monitoring Application. Sensors 2013, 13, 1787–1835. [CrossRef] [PubMed] Zhong, X.; Chan, H.; Rogers, T.J.; Rosenberg, C.P.; Coyle, E.J. The development and eStadium testbeds for research and development of wireless services for large-scale sports venues. In Proceedings of the IEEE 2nd International Conference on Testbeds and Research Infrastructures for the Development of Networks and Communities, Barcelona, Spain, 1–3 March 2006. Chen, C.; Wei, B.; Zhang, F. Design and realization of wireless environmental monitoring system based on ZigBee for stadium. In Proceedings of the IEEE International Conference on Future Computer Science and Education (ICFCSE), Xi’an, China, 20–21 August 2011; pp. 149–151. Cheng, L.; Hailes, S. Analysis of Wireless Inertial Sensing for Athlete Coaching Support. In Proceedings of the IEEE Global Telecommunications Conference, New Orleans, LA, USA, 30 November–4 December 2008; pp. 1–5. Wei, P.; Morey, B.; Dyson, T.; McMahon, N.; Hsu, Y.-Y.; Gazman, S.; Klinker, L.; Ives, B.; Dowling, K.; Rafferty, C. A conformal sensor for wireless sweat level monitoring. In Proceedings of the IEEE Sensors, Baltimore, MD, USA, 3–6 November 2013; pp. 1–4. Tao, W.; Liu, T.; Zheng, R.; Feng, H. Gait Analysis Using Wearable Sensors. Sensors 2012, 12, 2255–2283. [CrossRef] [PubMed] Chi, E.H. Introducing wearable force sensors in martial arts. IEEE Pervasive Comput. 2005, 4, 47–53. [CrossRef] Heinz, E.A.; Kunze, K.S.; Gruber, M.; Bannach, D.; Lukowicz, P. Using Wearable Sensors for Real-Time Recognition Tasks in Games of Martial Arts—An Initial Experiment. In Proceedings of the IEEE Symposium on Computational Intelligence and Games, Reno, NV, USA, 22–24 May 2006; pp. 98–102. Shaw, D.; Kavanal, B.K. Development of a multiple regression equation to predict Judo performance with the help of selected structural and body composition variables. In Proceedings of the IEEE First Regional Conference on Engineering in Medicine and Biology Society, 1995 and 14th Conference of the Biomedical Engineering Society of India, New Delhi, India, 15–18 February 1995; pp. 3/97–3/98. Baumann, J.; Wessel, R.; Kruger, B.; Weber, A. Action graph a versatile data structure for action recognition. In Proceedings of the International Conference on Computer Graphics Theory and Applications (GRAPP), Lisbon, Portugal, 5–8 January 2014; pp. 1–10.

Sensors 2016, 16, 1247

18.

19.

20.

21.

22. 23.

24.

25.

26.

27.

28.

29. 30.

31. 32. 33.

25 of 25

Lopez-Iturri, P.; Aguirre, E.; Azpilicueta, L.; Astrain, J.J.; Villadangos, J.; Falcone, F. Radio Characterization for ISM 2.4 GHz Wireless Sensor Networks for Judo Monitoring Applications. Sensors 2014, 14, 24004–24028. [CrossRef] [PubMed] Zulkifli, N.S.A.; Harun, F.K.C.; Azahar, N.S. XBee Wireless Sensor Networks for Heart Rate Monitoring in Sport Training. In Proceedings of the International Conference on Biomedical Engineering, Penang, Malaysia, 27–28 February 2012; pp. 441–444. Gong, M.; Zhang, L.; Ding, Z.; Dong, F.; Wang, L. Research and development of swimming training information system based on ZigBee technology. In Proceedings of the International Conference on Systems and Informatics, Yantai, China, 19–20 May 2012; pp. 966–970. Aguirre, E.; Arpon, J.; Azpilicueta, L.; de Miguel Bilbao, S.; Ramos, V.; Falcone, F.J. Evaluation of electromagnetic dosimetry of wireless systems in complex indoor scenarios with human body interaction. Prog. Electromagn. Res. B 2012, 43, 189–209. [CrossRef] Iskander, M.F.; Yun, Z. Propagation prediction models for wireless communication systems. IEEE Trans. Microw. Theory Tech. 2002, 50, 662–673. [CrossRef] Azpilicueta, L.; Rawat, M.; Rawat, K.; Ghannouchi, F.; Falcone, F. Convergence Analysis in Deterministic 3D Ray Launching Radio Channel Estimation in Complex Environments. Appl. Comput. Electromagn. Soc. J. 2014, 29, 256–270. Aguirre, E.; Lopez-Iturri, P.; Azpilicueta, L.; Astrain, J.J.; Villadangos, J.; Falcone, F. Analysis of Wireless Sensor Network Topology and Estimation of Optimal Network Deployment by Deterministic Radio Channel Characterization. Sensors 2015, 15, 3766–3788. [CrossRef] [PubMed] Aguirre, E.; López, P.; Azpilicueta, L.; Arpón, J.; Falcone, F. Characterization and Consideration of Topological Impact of Wireless Propagation in a Commercial Aircraft Environment. IEEE Antennas Propag. Mag. 2013, 55, 240–258. López-Iturri, P.; Nazábal, J.A.; Azpilicueta, L.; Rodriguez, P.; Beruete, M.; Fernández-Valdivielso, C.; Falcone, F. Impact of High Power Interference Sources in Planning and Deployment of Wireless Sensor Networks and Devices in the 2.4 GHz Frequency Band in Heterogeneous Environments. Sensors 2012, 12, 15689–15708. [CrossRef] [PubMed] Azpilicueta, L.; López-Iturri, P.; Aguirre, E.; Falcone, F. An accurate UTD Extension to a Ray Launching Algorithm for the Analysis of Complex Indoor Radio Environments. J. Electromagn. Waves Appl. 2016, 30, 43–60. [CrossRef] Azpilicueta, L.; Rawat, M.; Rawat, K.; Ghannouchi, F.; Falcone, F. A Ray Launching-Neural Network Approach for Radio Wave Propagation Analysis in Complex Indoor Environments. IEEE Trans. Antennas Propag. 2014, 62, 2777–2786. Azpilicueta, L.; Falcone, F.; Janaswamy, R. A Hybrid Ray Launching-Diffusion Equation Approach for Propagation Prediction in Complex Indoor Environments. IEEE Antennas Wirel. Propag. Lett. 2016. [CrossRef] Casino, F.; Azpilicueta, L.; Lopez-Iturri, P.; Aguirre, E.; Falcone, F.; Solanas, A. Hybrid-based Optimization of Wireless Channel Characterization for Health Services in Medical Complex Environments. In Proceedings of the 6th International Conference on Information, Intelligence, Systems and Applications (IISA 2015), Ionian University, Corfu, Greece, 6–8 July 2015. Komarov, K.K. Handbook of Dielectric and Thermal Properties of Materials at Microwave Frequencies; Artech House: Boston, MA, USA; London, UK, 2012. Aguirre, E.; Led, S.; Lopez-Iturri, P.; Azpilicueta, L.; Serrano, L.; Falcone, F. Implementation of Context Aware e-Health Environments Based on Social Sensor Networks. Sensors 2016, 16, 310. [CrossRef] [PubMed] Waspmote Datasheet. Available online: http://www.libelium.com/development/waspmote/ documentation/waspmote-datasheet/ (accessed on 5 August 2016). © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

Implementation and Analysis of ISM 2.4 GHz Wireless Sensor Network Systems in Judo Training Venues.

In this work, the performance of ISM 2.4 GHz Wireless Sensor Networks (WSNs) deployed in judo training venues is analyzed. Judo is a very popular mart...
13MB Sizes 0 Downloads 10 Views