A Real-time Vital Data Collection System for a Group of Persons during

a Variety of Sporting Activities in a Large Outdoor Ground

Shinsuke Hara

, Takunori Shimazaki

, Takuma Hamagami

, Yasutaka Kawamoto

, Hiroyuki Yomo

Ryusuke Miyamoto

, Hiroyuki Okuhata

and Fumie Ono

Graduate School of Engineering, Osaka City University, Sugimoto-Cho, Osaka-Shi, Osaka, Japan

Oki Electric Industry Co., Ltd., Bingo-Machi, Osaka-Shi, Japan

Graduate School of Science Engineering, Kansai University, Yamate-Cho, Suita-Shi, Osaka, Japan

School of Science and Technology, Meiji University, Higashi-Mita, Kawasaki-Shi, Kanagawa, Japan

Soliton Systems K.K., Kakuda-Cho, Osaka-Shi, Osaka, Japan

Wireless Networks Research Center, National Institute of Information and Communications Technology, Hikarino-Oka,

Yokosuka-Shi, Kanagawa, Japan

yomo@kansai-u.ac.jp, miya@cs.meiji.ac.jp, hiroyuki.okuhata@soliton.co.jp, fumie@nict.go.jp

Keywords:

Real-time Vital Data Collection, Heart Rate, VO2, Location, Wireless Multi-hop Networking, A Variety of

Sports Activities, A Group of Persons, A Large Outdoor Ground.

Abstract:

We have been developing a wireless vital data collection system named “AccuWiSe,” which is workable for

a group of persons during a variety of sporting activities in a large outdoor ground in real-time and reliably.

Using the second-prototype system, we have conducted an experiment on the 6th of March 2019, where

involving 50 subjects, we have successfully collected vital data from 18 subjects making a variety of sporting

activities in a sports ground with size of 60m×90m, in data collection rate of 94.9%, once in 2sec regularly,

and for 45min continuously. This paper introduces AccuWiSe and demonstrates the experimental results.

1 INTRODUCTION

Reliable and real-time vital signs monitoring is re-

quired to promote health and prevent disease/injury

for persons during sporting activity. For professional

and amateur athletes, it can be also used for evidence-

based physical training to improvetheir performances

and identify their talents. We have been develop-

ing a real-time vital signs monitoring system for a

group of persons during a variety of sporting activi-

ties in a large outdoor ground and have named it “Ac-

curate Wireless vital Sensing system (AccuWiSe).”

In the development of AccuWiSe, using the ﬁrst-

prototype wireless vital sensor nodes (VSNs) and de-

signed wireless networking protocol, we conducted

an experiment on the 26th of February 2018 (Ham-

agami et al., 2018), where at a data collection node

(DCN), we collected the data

• of heart rate (HR), VO2, bodysurface temperature

and humidity, and air temperature and humidity,

• from wireless vital sensor nodes (VSNs) put to the

forearms of 22 footballers,

• once in 1 sec during 4×15min football matches,

• in an outdoor football ﬁeld of 55m×90m,

• using the wireless communication tool with data

rate of 100kbps and transmission power of 20mW

in the 920MHz band,

• in data collection rate of 97.9%.

The experiment was very successful, but it had

three major drawbacks; the ﬁrst-prototypeVSNs were

large, the networking protocol was able to accommo-

date only up to 25 VSNs, and the performance was

evaluated only in a single exercise of football match

where footballers are likely to randomly spread in the

entire ground. According to the feedback and re-

ﬂection on the system development and the method

of performance evaluation, we have developed the

second-prototype system and have conducted an ex-

periment on the 6th of March 2019.

In this paper, we introduce AccuWise focusing on

its wireless networking technique and show the latest

experimental results using its second-prototype sys-

tem.

138

Hara, S., Shimazaki, T., Hamagami, T., Kawamoto, Y., Yomo, H., Miyamoto, R., Okuhata, H. and Ono, F.

A Real-time Vital Data Collection System for a Group of Persons during a Variety of Sporting Activities in a Large Outdoor Ground.

DOI: 10.5220/0008066401380145

In Proceedings of the 7th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2019), pages 138-145

ISBN: 978-989-758-383-4

This paper is organized as follows. Section 2

explains the reason to select the 920MHz band and

outlines our multi-hop networking protocol in Ac-

cuWiSe. Section 3 shows the speciﬁcations of the

second-prototype system, focusing on its difference

from the ﬁrst-prototype system. Section 4 demon-

strates and discusses the experimental results. Finally,

Section 5 concludes the paper.

2 WIRELESS NETWORKING

TECHNIQUE

2.1 Selection of Frequency Band

We conducted an experiment on real-time vital data

collection for 22 footballers during a match in an out-

door football ground (Hara et al., 2013). The main

purpose of the experiment was to compare the per-

formance between wireless communication tools in

the 920MHz and 2.4GHz bands, so we implemented

VSNs which can transmit vital data with the same

packet length in the same timings in both frequency

bands, where the wireless communication tools were

compliant with the ARIB STD-T108 (ARIB, 2011)

and the IEEE 802.15.4 standard (IEEE, 2015) with

transmission powers of 20mW and 10mW in the

920MHz and 2.4GHz bands, respectively.

As shown in Figure 1 (a), we put the VSNs to the

back waist positions of all footballers and also placed

6 data forwarding nodes (DFNs) around the ground.

The VSNs broadcast their own sensed data to the

DFNs once in 10 sec and the DFNs forwarded their

received data to a single data collection node (DCN)

directly or through other DFNs. Figure 1 (b) shows

the performanceon the packet success rate obtainedin

the experiment. For the wireless communication tool

in the 2.4GHz band, even when using 6 DFNs around

the ground, the packet loss rate cannot be more than

than 75%, on the other hand, for the wireless commu-

nication tool in the 920MHz band, when using 1 or

2 DFNs around the ground, the packet loss rate can

be more than 95%. This is because the signal in the

920MHz band has a longer transmittable range and

is less vulnerable to fading and blocking by human

body.

Based on the experimental result, we decided

to select the wireless communication tool in the

920MHz band as the one suitable for the real-time

vital data collection from a group of persons spread

in a large outdoor ground. Note that some real-

time vital data collection systems have been com-

mercially available in the market, but they operate in

2 3

Vital sensor node

Data collection node

Data forwarding node

(a) Layout

(b) Performance

1 2 3 4 5 6

Number of data forwarding nodes

100

Packet success rate [%]

920 MHz band

2.4 GHz band

Antenna height of

data forwarding node

=1m

Figure 1: Layout and performance for the broadcast/

forward-based vital data collection.

the 2.4GHz industrial, scientiﬁc and medical (ISM)

band such as WiFi (https://www.wi-ﬁ.org/), Bluetooth

(https://www.bluetooth.com/) or Bluetooth Low En-

ergy (https://www.bluetooth.com/). We imagine that

a lot of packet losses would occur in them so the lost

data might be replaced by the ones previously suc-

cessfully received, but physical or technical trainers

could not notice the fact.

2.2 Multihop Networking Protocol

We decided to use the wireless communication tool

in the 920MHz, but when exercisers, namely, VSNs,

spread in a large ground, it cannot directly connect

them to a DCN, even though it has a longer transmit-

table range and less vulnerability against fading and

blocking. One approach could be to place DFNs or

relay nodes in the ground, but through the experiment,

we noticed it is troublesome; there is often something

wrong with DFN and we need to replace the battery

of DFN. Therefore, as shown in Figure 2, we took an

approach of multihop data collection through VSNs.

A Real-time Vital Data Collection System for a Group of Persons during a Variety of Sporting Activities in a Large Outdoor Ground

139

Vital sensor node

Data collection node

Data collection route

Figure 2: Multihop data collection through VSNs.

We evaluated some ready-made multihop net-

working protocols such as Ad hoc On-Demand Dis-

tance Vector (AODV) (Perkins, Belding-Royer, &

Das, 2003), but the experimental packet success rate

was terribly low. According to the AODV, VSNs

always exchange control packets to discover their

neighbors and update their routing tables. When a

VSN generates or receives a packet to transmit for a

DCN, the VSN forwards the packet to its parent VSN

according to its own routing table. When VSNs are

stationary, the AODV works well, but when they are

in motion, in other words, the network topology is dy-

namically changing, the AODV does not work well,

since when a VSN generates or receives a packet to

transmit, its routing table has been already old and

invalid. Therefore, we decided to design a multihop

network protocol valid for our application.

We designed a ﬂooding/time division multiple ac-

cess (TDMA) protocol (Hamagami et al., 2018), the

principle of which is a superframe-by-superframeDy-

namic Source Routing (DSR) (Johnson, Hu, & Malts,

2007). According to the ﬂooding/TDMA protocol,

the system operates on two stages such as “pairing

stage” and “collection stage.”

The pairing stage starts when the switches of the

VSNs are on. Each VSN tries to transmit a pairing

request packet to a DCN, and when the DCN success-

fully receives the packet, the DCN transmits a pairing

reply packet to the VSN, assigning an identiﬁer num-

ber (ID) to it. In the ﬁrst-prototype system, an distinct

ID, which corresponds to the TDMA time slot ID, is

assigned to each VSN. Therefore, as explained later,

any VSN can freely and solely begin to transmit its

data frame in its assigned TDMA time slot as its “slot

owner.” When all the VSNs have received their pair-

ing reply packets, network association completes.

Figure 3 shows the superframe structure of the

ﬂooding/TDMA protocol used in the collection stage,

which is composed of a ﬂooding period and a data

Flooding period Data collection period

150msec

2.96msec 77.04msec 70msec

Slot 1 Slot 2 Slot 3 Slot 1

33msec

Slot 2

...

Slot m

...

Slot 22

726msec

1,000msec

Inactive period

Superframe

124msec

7.36msec

Frame 1 Frame 2 Frame 3

33msec

2.80msec 0.16msec

Header CRC bits

3.04msec 4.16msec 0.16msec

Header Payload CRC bits

Beacon

Figure 3: Structure of the superframe.

1 (2)

1 (4)

D (1)

2 (3)

Time

2 1 1 D 3 2 2 1 1 D 4 1 1 D1 D

Uplink timeslot 1 Uplink timeslot 2 Uplink timeslot 3 Uplink timeslot 4

DCN

VSN 1

VSN 2

VSN 4

VSN 3

1 (2)

1 (4)

D (1)

2 (3)

(a) Flooding period

Time

DCN

VSN 1

VSN 2

VSN 4

VSN 3

2 1

4 1

1 D

3 2

(b) Data collection period

TX node attributes

Hop count

CSMA period

TX node attributes

Data and data size

Figure 4: Operation of ﬂooding and data collection.

collection period. The role of the former is to select a

suitable parent VSN for each VSN whereas that of the

latter is to forward a data frame from each VSN to the

DCN through other VSNs without frame collisions.

Here, taking into consideration of the ground size and

transmittable range of VSN, we limit the number of

hops to 3 in the time slot for each VSN.

The communication among the DCN and VSNs is

divided into a series of superframes, and the collec-

tion stage starts when the DCN broadcasts a beacon.

In the ﬂooding period, only the DCN can initiate the

transmission of a beacon. When any VSN receives

a beacon broadcast by the DCN or another VSN, it

re-broadcasts a beacon showing its own node ID and

its own hop count in the beacon. In addition, when

any VSN receives multiple beacons from other VSNs

within a certain period, it measures the receivedsignal

strength (RSS) for each received beacon, and it mem-

orizes the node ID which gives the largest RSS as its

parent VSN. This is the reason why the second and

third slots are longer than the ﬁrst slot. The maximum

number of hops is set to three in the the ﬁrst-prototype

system, so the ﬂooding period is divided into three

time slots.

In the data collection period, each TDMA time

slot has been assigned to a distinct VSN as its slot

owner in the pairing stage, so any VSN can initiate its

data frame transmission in its assigned slot. In addi-

icSPORTS 2019 - 7th International Conference on Sport Sciences Research and Technology Support

140

tion, when a VSN receives a data frame, it unicasts

the frame to its parent node.

Finally, Figure 4 shows one operational example

of the ﬂooding and data collection periods.

3 SECOND-PROTOTYPE

SYSTEM

3.1 Implementation and Experiment of

the First-prototype System

One of problems in the development of networking

protocol is that there is no repeatability in experiments

using real subjects in the sense that we cannot exper-

imentally compare the performances among different

protocols at the same time. Therefore, we developed

a network simulator which is composed of a channel

model set and a mobility model set, and evaluating

the performances of different networking protocols

changing their adjustable parameters repeatedly in the

same environmental situation, we determined the pa-

rameters of the ﬂooding/TDMA protocol (Hara et al.,

2018), which are shown in Figure 3. And ﬁnally,

we evaluated the performance of the ﬁrst-prototype

system in experiments involving 22 footballers dur-

ing 4×15min matches once indoors (ﬁrst) and twice

outdoors (second and third) in 2017 and 2018 (Ham-

agami et al., 2018). The speciﬁcations and experi-

mental results on the ﬁrst-prototype system are sum-

marized as follows:

• VSN: Size=45mm×44mm×15mm, weight=31g,

power consumption=90mA,

• Packet success rate: 98.3% (ﬁrst indoor, back

waist), 92.6% (second outdoor, back waist),

97.8% (second outdoor, forearm), 92.1% (second

outdoor, calf), and 97.9% (third outdoor, fore-

arm).

3.2 Implementation of the Second

Prototype System

We evaluated the performance of the ﬁrst-prototype

system by the experiments and obtained the success-

ful results, but the system and the method of perfor-

mance evaluation had three major drawbacks. The

ﬁrst one is that the VSN was composed of two pieces

of circuit boards, one of which was equipped with

a micro controller unit (MPU) handling several vital

sensor devices and the other of which was a ready-

made transceiver module also containing an MPU. As

a result, the size of the VSN was larger. In addition,

CPU

EFM32GG295F1024

HTU21D(F)

Body surface

thermohydrometer

CBT Sensor

HTU21D(F)

Air

thermohydrometer

2x4Bytes, 1Hz

CAM-M8Q

GPS

NMEA (GLL)

Location sensor

45Bytes, 1HzI2C UART

Wireless transceiver

Si4461

Tri-axial

Accelerometer

MPU-9250

x y z

VO2 Sensor

Normal PPG Sensor

Color sensor

BH1792

(Green LED)

HR Sensor

MA Sensor

Color sensor

BH1792

(Green LED)

3x4Bytes, 8Hz 2x2Bytes, 32Hz

I2C

920MHz band

20mW

100kbps

48MHz

SPI

Figure 5: Block diagram of the second-prototype VSN.

(a) Vital sensor node (b) Wearing position

Battery

Dipole

Antenna

GPS antenna

Vital sensor node

Figure 6: Photos of vital sensor node and its wearing posi-

tion (forearm).

20hPa 30hPa 50hPa 70hPa

RMSPE [%]

PPG

PPG/MAC/OR

PPG/MAC/PFT

(Loose) (Moderate) (Tight) (Very tight)

Wearing pressure

Forearm

Figure 7: Performance comparison between OR and PFT.

a passive antenna was used for the GPS receiver, so

it did not work at all. The second one is that the de-

signed superframe was able to accommodate up to 25

TDMA time slots, in other words, the system was able

to accommodate up to 25 persons for vital sensing.

Finally, the third one is that we evaluated the perfor-

mance in a single sporting activity of football, where

footballers are likely to randomly spread in the entire

ground.

According to the feedback and reﬂection on

the system development and the method of perfor-

mance evaluation, we have implemented the second-

prototype system.

A Real-time Vital Data Collection System for a Group of Persons during a Variety of Sporting Activities in a Large Outdoor Ground

141

Table 1: Payload format.

Sequence Number 2Bytes

HR 1Byte

Body surface temperature/humidity 4Bytes

Air temperature/humidity 4Bytes

VO2 1Byte

GPS longitude/latitude 4Bytes

Acceleration 12Bytes

Reserved 2Bytes

3.2.1 VSN

Figure 5 shows the block diagram of the second-

prototype VSN. A single MPU based on ARM

Cortex-M3 handles all the devices, so the VSN is

composed of a single circuit board. In addition, some

sensor components are changed, so the power con-

sumption is reduced to 70mA (22.2% reduction). On

the other hand, we adopt a larger battery (CLB3032)

and a larger active antenna (15dB gain) for GPS re-

ceiver, so the size of the VSN is not so reduced as

48mm×38×15mm (7.9% reduction) and its weight is

32g including the battery and antenna.

Regarding the wearing position of VSN, we de-

cide to put the VSN to the forearm of person, since

the forearm gives the highest packet success rate in

the experimental results by the ﬁrst-prototype system.

Figures 6 (a) and (b) show the photos of the second-

prototype VSN and its wearing position at a person,

respectively.

Furthermore, regarding the signal process for HR

calculation, we replace the OR technique by the peak

frequency tracking (PFT) technique (Zhang, Pi, &

Liu, 2015). Figure 7 compares the HR during sport-

ing activities with different intensities between the

OR and PFT techniques, which are obtained by ex-

periments involving 13 subjects. We can conﬁrm the

improvement by the PFT in all the range of wearing

pressure.

Finally, Table 1 shows the payload format in the

data frame.

3.2.2 Multihop Networking Protocol

The second-prototype system is based on the same

ﬂooding/TDMA networking protocol as in the ﬁrst-

prototype system, but the superframe structure is

changed as an extended superframe structure. Fig-

ure 8 shows the structure of the extended superframe,

where each extended superframe is composed of n su-

perframes. To be able to accommodate up to 30 VSN

in a superframe, the preamble and payload are con-

densed while keeping the amount of information the

same as for the ﬁrst-prototype system. According to

Flooding period Data collection period

380msec

4.16msec 77.04msec 70msec

Slot 1 Slot 2 Slot 3 Slot 1

33msec

Slot 2

...

Slot m

...

Slot 30

600msec

1,000msec

Inactive period

Superframe

20msec

5.68msec

Frame 1 Frame 2 Frame 3

20msec

3.12msec 2.40msec 0.16msec

Header Payload CRC bits

nx1,000msec

Extended superframe

1,000msec

Superframe 1

...

Superframe 2 Superframe n

Beacon

2.40msec 1.60msec 0.16msec

Header Payload CRC bits

Figure 8: Structure of the extended superframe.

the pairing stage for the extended superframe struc-

ture, a DCN ﬁrst randomly divides a whole VSNs into

n groups, and then assigns a distinct TDMA time slot

in the nth superframe to each VSN beloging to the nth

group.

Table 2: Detail on the experiment.

Date 6th of March, 2019

Place Kita-Yamoto baseball

ground

Area 60m×90m

Number of subjects 50

Number of VSNs 18

Number of DCN 1

Sampling rate 8samples/sec

Transmission

frequency band

920MHz band

Wireless transmis-

sion standard

ARIB-T108

Transmission

power

20mW

Transmission rate 100kbps

Duration 45min

4 EXPERIMENTAL RESULTS

AND DISCUSSIONS

Using the second-prototype system, we have con-

ducted an experiment. To involve 50 subjects in the

experiment, we set the parameter of extended super-

frame n=2, so the system was able to accommodata

up to 60 VSNs. We had a plan to implement 50 VSNs

and 1 DCN before the date of the experiment, but

there was something wrong with the batteries and cir-

cuits of VSNs, so we were able to ﬁnish implement-

ing18 VSNs. Therefore, in the experiment, we put

the operable VSNs to the forearms of 18 subjects ran-

domly selected out of 50 subjects. The DCN could as-

icSPORTS 2019 - 7th International Conference on Sport Sciences Research and Technology Support

142

Antenna

Data collection node

(d) Warming-up exercises (e) Mini-football exercise (f) 50m footrace

(a) Data collection node (b) Panoramic view

Figure 9: Photos of the experiment.

0 5 10 15 20 30 40 45

60 80 100 120 140 160 180 200

25 35

Time [min]

Heart rate [beats per minute]

220

Warmig-up exercises

Mini-football exerciseEntry jogging

Random jogging

50m footrace

Gathering

0 5 10 15 20 30 40 45

0 5 10 15

20 25 30

25 35

Time [min]

VO2 [ml/kg/min]

Warmig-up exercises

Mini-football exerciseEntry jogging

Random jogging

50m footrace

Gathering

(a) HR

(d) VO2

0 5 10 15 20 30 40 45

32 33 34 35 36 37 38 39

25 35

Time [min]

Temperature [degrees C]

Mini-football exerciseEntry jogging 50m footrace

Warmig-up exercises Random jogging Gathering

(b) Body surface temperature

(e) Air temperature

0 5 10 15 20 30 40 45

32 33 34 35 36 37 38 39

25 35

Time [min]

Temperature [degrees C]

Mini-football exerciseEntry jogging 50m footrace

Warmig-up exercises Random jogging Gathering

(e) Air humidity

0 5 10 15 20 30 40 45

0 10 20 30 40 50 60 70

25 35

Time [min]

Humidity [%]

Mini-football exerciseEntry jogging 50m footrace

Warmig-up exercises Random jogging Gathering

90 100

0 5 10 15 20 30 40 45

0 10 20 30 40 50 60 70

25 35

Time [min]

Humidity [%]

Mini-football exerciseEntry jogging 50m footrace

Warmig-up exercises Random jogging Gathering

90 100

Figure 10: Sensed vital data for all the 18 subjects.

sign the TDMA time slots in one of the 2 superframes

to all the VSNs, but it randomly assigned the TDMA

time slots of the 2 superframes, that is, 9 VSNs in one

superframe and another 9 VSNs in the other super-

frame. Table 2 shows the details on the experiment.

In addition, we have designed the protocol of the

experiment so as to include a variety of sporting ac-

tivities, such as uniformly spreading in the entire area,

gathering in a localized area, with low and high mo-

bility, with random and coordinated mobility and so

on. Table 3 shows the protocol of the experiment for

45min in total period, and Figure 9 shows the photos

of experiment.

Figures 10 (a)-(f) show the temporal variations on

the HR, VO2, body surface temperature/humidity and

air temperature/humidity for all the 18 subjects, re-

spectively. We can see from these ﬁgures that the

second-prototype system can collect the vital data

from all the 18 subjects regularly for 45min. For the

HR sensing, we put the Holter monitors to 2 subjects,

A Real-time Vital Data Collection System for a Group of Persons during a Variety of Sporting Activities in a Large Outdoor Ground

143

Table 3: Protocol of the experiment.

Entry jogging 4min

Gathering into 1 group 1min

Spreading 1min

Warming-up exercises 4min

Gathering 1min

Spreading into 4 groups 2min

Mini-football exercise 10min

Gathering into 4 groups 1min

Resting 2min

Random jogging 3min

Gathering into 4 groups 1min

Dividing into 12 groups 1min

Random jogging 3min

Gathering into 1 group 1min

Resting 2min

50m footrace 6min

Gathering into 1 group 1min

Resting 1min

and we conﬁrmed that the HR sensors worked accu-

rately for them. For the temperature and humidity

sensing, the VSN senses the temperature and humid-

ity of the air in the case of VSN through a hole, so

they may be affected by those of body surface. How-

ever, the air temperature tends to gradually decrease,

whereas the body surface temperature tends to keep

ﬂat. On the other hand, the body surface humidity

tends to earlier saturate (reach 100%) due to sweat by

sporting activities, whereas the air humidity tends to

gradually increase. Indeed, it began raining around

the end of the experiment.

Figure 11 shows the packet success rates for the

active 18 VSNs. All the packet success rates are satis-

factorily high, the average=94.9%, so we can see that

the ﬂooding/TDMA protocol using the extended su-

perframe structure works effectively. In addition, Fig-

ure 12 shows the distribution on the number of hops.

It is interesting to see that even for the large ground,

around 90% of packets reach the DCN directly, but

the multihop data collection is essential to make the

packet success rate higher.

Finally, Figure 13 focuses the sensed vital data

and location for a subject. It is quite natural to see

that the HR temporal variation is likely to be synchro-

nized with that of VO2, but there is some disagree-

ment between them around at the beginning of the

experiment. We could imagine that it is because the

subject was not tired when the experiment started.

1 2 3 4 5 18

VSN number

Packet success rate [%]

100

7 8 9 10 11 12 13 14 15 16 17

Average=94.9%

Figure 11: Packet success rate for all the 18 subjects.

3 hops (0.2%)

1 hop (90.2%)

2 hop2 (9.6%)

Figure 12: Distribution on the number of hops.

5 CONCLUSIONS

In this paper, after outlining the vital sensing and

wireless networking techniques which we have been

improving, we showed the latest experimental result

on the real-time vital data collection systems for a

group of persons during sporting activities in a large

outdoor ground. Because of some troubles in imple-

mented vital sensor nodes and batteries, the number

of operable vital sensor nodes used in the experiment

was not satisfactorily high but we conﬁrmed that the

second-prototypesystem works effectively for 18 per-

sons in a group of 50 persons during a variety of sport-

ing activities for 45min.

We have not developed a technique to estimate

core body temperature using information by wearable

vital sensors, so one of our future works includes its

development. In addition, we need to accommodate

up to 150 persons while improving the packet success

rate up to 99.0%, so another of our future works in-

cludes the design of a more efﬁcient ﬂooding/TDMA

networking protocol based on the extended super-

frame structure.

ACKNOWLEDGMENTS

The authors have disclosed no other potential con-

ﬂicts of interest (COI). In addition, this research was

approved by the institutional review board of one of

the author’s organizations.

icSPORTS 2019 - 7th International Conference on Sport Sciences Research and Technology Support

144

0 5 10 15 20 30 40 45

60 80 100 120 140 160 180

25 35

Time [min]

Heart rate [beats per minute]

200

Warmig-up exercises

Mini-football exercise

Entry jogging

Random jogging

50m footrace

Gathering

25 20

15 10 5

VO2 [ml/kg/min]

035 30

Heart rate

VO2

(a) HR and VO2

0 5 10 15 20 30 40 45

30 31 32 33 35 37 39

25 35

Time [min]

Heart rate [beats per minute]

Warmig-up exercises

Mini-football exercise

Entry jogging

Random jogging

50m footrace

Gathering

80 70 60 40 20

Humidity [%]

0100 50 30 1090

34 36 38

Body surface temperature

Body surface humidity

Air temperature

Air humidity

(b) Temperature and humidity

Entr

y jo

gging

Warmig-up exercises

Mini-football exercise

50m footrace

Ran

m jogging

0 20 40 60 80 100 120 140 [m]

10 20 30 40 50 60 70 90 [m]800

Figure 13: Sensed vital data and location for a subject.

REFERENCES

Hamagami, T. et al., 2018. Wireless Multi-Hop Networking

for a Group of Exercisers Spread in a Sports Ground,

Proc. IEEE Healthcom 2018, pp. 1-6.

Hara, S. et al., 2013. Development of a Real-Time Vital

Data Collection System from Players during a Foot-

ball Game, Proc. IEEE Healthcom 2013, pp. 1-5.

Association of Radio Industries and Businesses, 2011.

ARIB STD -T108, version 1.0.

IEEE, 2015. IEEE Standard for Low-Rate Wireless Net-

works 802.15.4-2015.

Perkins, C., Belding-Royer, E., & Das., S., 2003. Ad hoc

On-Demand Distance Vector (AODV) Routing, IETF,

RFC 3561.

Johnson, D., Hu, Y., & Malts, D., 2007. The Dynamic

Source Routing Protocol (DSR) for Mobile Ad Hoc

Networks for IPv4, IETF, RFC 4728.

Hara, S. et al., 2018. Performance prediction of wireless

multi-hop networks using stored data sets for persons

during exercises, Proc. ICT 2018, pp. 1-6.

Zhang, Z., Pi, Z., & Liu, B., 2015. TROIKA: A General

Framework for Heart Rate Monitoring Using Wrist-

Type Photoplethysmographic Signals During Inten-

sive Physical Exercise, EEE Trans. on Biomed. Eng.,

vol. 62, no. 2, pp. 522-531.

A Real-time Vital Data Collection System for a Group of Persons during a Variety of Sporting Activities in a Large Outdoor Ground

145