Development of WiFi Mesh Infrastructure for Internet of Things Applications

Abstract

This paper described the design of WiFi mesh infrastructure in order to improve the performance of an ad hoc network, WLAN, and WMAN so that the interconnection becomes easier and cheaper. The built system consisting of some WiFi routers were configured to form a WiFi mesh network so that it can build the Internet cloud. The operating system used in this research is quick mesh project (QMP). In addition, the WiFi node as a client developed for the internet of things (IoT) applications. These nodes using message queuing protocol telemetry transport (MQTT) based publish and subscribe. WiFi mesh infrastructure is working well with the average value of the measured throughput was 110.5 kbps.

Full text

Procedia Engineering 170 ( 2017 ) 332 – 337
1877-7058 © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the Engineering Physics International Conference 2016
doi: 10.1016/j.proeng.2017.03.045
ScienceDirect
Available online at www.sciencedirect.com

Engineering Physics International Conference, EPIC 2016
Development of WiFi Mesh Infrastructure for Internet of Things
applications
Rifki Muhendra, Aditya Rinaldi, Maman Budimana*, Khairurrijal
Department of Physics, Faculty of Mathematics and Natural Sciences,
Institut Teknologi Bandung, Jalan Ganesa 10, Bandung 40132
Abstract
This paper described the design of WiFi mesh infrastructure in order to improve the performance of an ad hoc network, WLAN, and WMAN
so that the interconnection becomes easier and cheaper. The built system consisting of some WiFi routers were configured to form a WiFi
mesh network so that it can build the Internet cloud. The operating system used in this research is quick mesh project (QMP). In addition, the
WiFi node as a client developed for the internet of things (IoT) applications. These nodes using message queuing protocol telemetry transport
(MQTT) based publish and subscribe. WiFi mesh infrastructure is working well with the average value of the measured throughput was 110.5
kbps.
© 2016 The Authors. Published by Elsevier Ltd.
Keywords: WiFi mesh, QMP, MQTT, IoT
1. Introduction
Wireless mesh network (WMN) is a communications network made up of radio nodes organized in a mesh topology. WMN
usually consist of mesh clients, mesh routers and gateways. Clients are electronic devices, embedded systems, and sensors that
can communicate with each other in a network. The router is an electronic device serves as a liaison between two or more
networks to carry data from one network to another. Gateway is an electronic device that bridges the network to the Internet [1].
When one node can no longer operate, the rest of the nodes in the WMN can still communicate with each other, either directly or
through one or more intermediate nodes.
Excellence on a WMN is the ease in adding new nodes without re-installation as a whole. WMN have been used for several
applications in the field of remote monitoring and control systems such as environmental monitoring [2], the water meter
infrastructure [3], automatic street light monitoring system [4], and Real-Time Indoor Carbon Dioxide Monitoring [5]. These
studies use a wireless module technology based physical layer such as Zigbee and radio frequency (RF) and thus require a
gateway to connect to the internet. This wireless mesh network has low security. Therefore, research that can establish a wireless
mesh network based on the latest technology and cover the lack of previous studies is important.
In this paper, we propose a WMN infrastructure based on WiFi technology that has the ease of installation, connectivity and
can be used in applications IoT. Wifi is a wireless communication technology-based on internet protocol (IP) that has the quality
of communication and better data security than the physical layer technology. In this study, we constructed mesh wifi router as
an Internet cloud and the wifi node as a client. Wifi Client designed small size, low power and easy to install with other devices.
Through this infrastructure, internet connectivity can reach significant level and support the IoT applications in various fields.

Corresponding author. Tel.: +62-22-2500834; fax: +62-22-2506452.
E-mail address:maman@fi.itb.ac.id
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the organizing committee of the Engineering Physics International Conference 2016

333
Rifki Muhendra et al. / Procedia Engineering 170 ( 2017 ) 332 – 337

2. Design of system
WiFi mesh infrastructure built is composed of WiFi client, WiFi routers, and gateways as shown in fig 1.
Fig. 1. WiFi mesh infrastructure
In the system that we have built, we designed a WiFi client used as a bridge that connects various kinds of electronic devices
to the mesh network. These devices can be a meter-meter, air-conditioning, lights and sensors. WiFi Client connected to the
internet via a WiFi router. Several adjacent WiFi routers will form a mesh topology that builds the internet cloud to the
surrounding area. One of the router to function as a gateway and connected to the internet via 3G / 4G, DSL, Ethernet, and
ISDN.
2.1 Design of WiFi client
WiFi client hardware design that has been built is shown in Figure 2. WiFi client has six main components, namely power
supply, a microcontroller (ATmega328), sensors / actuators, a watchdog timer (WDT), EEPROM and WiFi module (esp8266).
A power supply is an electronic device that supplies electrical energy to WiFi clients. The power supply has an input voltage of
220 VAC and an output voltage of 3.3 VDC. A microcontroller that we use is ATmega328. This type of microcontroller 8 bits
which have 32KB ISP flash memory with read-while-write capability, 1KB EEPROM, 2KB of SRAM, 23 general purpose I / O
lines, 32 general purpose registers work, three flexible time / counter with compare modes, interruptions internal and external,
programmable serial USART, 2-wire serial interface byte-oriented, SPI serial ports, 6-channel 10-bit A / D converter (8-
channels in TQFP and QFN package / MLF), watchdog programmable timer with internal oscillator and five software selectable
power saving mode. the device operates between 1.8 to 5.5 volts [6]. The microcontroller is a device that regulates all processing
logic on the client wifi like recording sensor data, communicates with the Internet and gives a command to the external device
connected to the client wifi. In addition, the microcontroller will perform data storage to the EEPROM when power loss.
WDT is an additional circuit on the client wifi which serves to restart the microcontroller when the failed process. WDT is
based timer 555. This timer provides a moment of time to the microcontroller to work and then will restart automatically.
Therefore, we added a simple program on the microcontroller to prevent this restart. WDT has been widely used by people
because it is simple and easy in its regulation
On the WiFi client, we use esp8266 as communication modules and connected in series with the microcontroller. Esp8266 is
radio-based protocol TCP / IP stack. Esp8266 works at a frequency of 2.4 GHz which can be integrated with sensors and other
electronic devices using GPIO. ESP8266 has been designed for mobile, wearable electronics and Internet of Things applications
with the aim of achieving the lowest power consumption with a combination of several proprietary techniques [7].

334 Rifki Muhendra et al. / Procedia Engineering 170 ( 2017 ) 332 – 337
Fig. 2. WiFi client
In communicating, Client WiFi uses MQTT messaging protocol. This is a messaging protocol designed for machine to
machine communication for small data-based "publish and subscribe" [8]. There are two main MQTT components are client and
broker. MQTT client may be a series of instrumentation, mobile devices and web applications that work published or subscribed
to a message. MQTT broker is an embedded computing platform on the internet that works to accommodate all of the published
data and deliver the data to the subscriber. A message that has been issued by the client can be subscribed by one or multiple
devices through a topic. The topic is a word that covers the data to be published. As long as the subscriber has the suitability
topic that will be subscribed, then the publisher and customer relationships will occur.
2.2 Development of Internet cloud
Internet cloud built using multiple WiFi routers one of which serves as a gateway. Hardware routers used is TP-LINK
MR3020. TP-LINK MR3020 using Atheros AR9331 400 MHz processor, 4 MB Flash Memory, Memory 32MB SDRAM, 1
Ethernet port 100 and works on 5 VDC 1.0 A. This router is a type of portable 3G / 4G which has a data transfer speed
150Mbps, small size, and low energy consumption. By connecting USB card 3G/4G router, Wi-Fi hotspot is instantly
established allowing users to share an Internet connection anywhere 3G/4G coverage available [9].
In the internet cloud builders, routers are implanted by the operating system that allows the establishment of a wireless mesh
network to extend the Internet uplink. The operating system used in this study is quick mesh project (QMP) based OpenWRT
Linux. As main routing protocol QMP uses BMX6, a destination-sequenced distance-vector protocol using UDP messages to
discover other nodes and disseminate node and routing information. Some extra features have been specially developed for
QMP, such as a smart gateway selection using IPIP tunnels or a short message plugin which permits to send arbitrary
information to other nodes, piggybacked by the protocol packages. BMX6 obtained some of the best results over other mesh
routing protocols tested in the Wireless Battle Mesh v6 celebrated in Aalborg (Denmark). Other important characteristics of
QMP are the native IPv6 and full auto-configuration support. Each node auto-configures its own IPv6 address based on a ULA4
prefix. IPv4 connections are enabled via tunnels over the ULA-based IPv6 network [10].
3. Results and Discussion
WiFi Client hardware that had been built can be seen in fig 5 (a). WiFi client is sized 4x4 cm with a source voltage of 3.3
VDC and can be planted a program through serial communication. In addition, the client WiFi is easy to install on electronic
devices, meters, and sensor circuit through I/O pins. To determine the client WiFi that has been built has a good performance,
we conducted the experiment of sending data to the internet. This data is room temperature was measured using a sensor LM35.
Microcontroller records the value of the room temperature every second. Those values are formatted into data packets and then
published to the MQTT broker on the topic "/ espeduino / temperature". Interface used to observe the temperature data is
MYMQTT as shown in Figure 5 (b). This interface is an android app that supports remote observation and control protocol
based MQTT. Subscribe topic with appropriate and good internet connectivity on android devices, data room temperature can be
displayed every second.


335
Rifki Muhendra et al. / Procedia Engineering 170 ( 2017 ) 332 – 337

(a) (b)
Fig. 3. Client WiFi IoT (a) hardware (b) interface
In addition to a WiFi client that has worked well, we also perform measurements of throughput and latency on WiFi mesh
network. Measurement throughput aims to determine the speed of the data sent to all terminals in the network. This
measurement is done by accessing a website by mobile devices in areas specified then a server will measure the speed of the
data transfer. Measurement throughput values were first carried out in the area around the gateway is located. Measurements
lasted for 4 minutes with 10 times the number of observations and the number of packets of data that is accessed varied. Results
of measurement throughput in the first area can be seen in table 1.
Table 1. measurement throughput values in the area around the gateway
Examination Min
(kbps)
Max
(kbps)
Average
(kbps)
1 43 159 101
2 34 132 83
3 35 142 88,5
4 47 187 117
5 60 203 131,5
6 63 198 130,5
7 58 176 117
8 65 210 136
9 38 134 86
10 46 183 114,5
Average 110,5
Table 1 shows the average value of the first area throughput is 110.5 kbps. The maximum value of the measured throughput is
210 kbps on the 8th observation. A minimum value of the measured throughput is 34 kbps on 2nd observation. Based on Table 1
it can be concluded that the first measurement area throughput speed internet access is good.
Measurement throughput values also do the outermost areas can still access the internet via WiFi mesh router. Measurements
lasted for 4 minutes with 10 times the number of observations and the number of packets of data that is accessed varied. Results
measurement throughput values in the outer region can be seen in Table 2.

336 Rifki Muhendra et al. / Procedia Engineering 170 ( 2017 ) 332 – 337
Table 2. measurement throughput in the outermost area
Examination Min
(kbps)
Max
(kbps)
Average
(kbps)
1 32 134 83
2 34 132 83
3 23 107 65
4 21 134 77,5
5 34 156 95
6 53 117 85
7 42 142 92
8 39 169 104
9 28 98 63
10 27 112 69,5
Average 81,7
Table 2 shows the average value of the first area throughput is 81.7 kbps. The maximum value of the measured throughput is
169 kbps on the 8th observation. A minimum value of the measured throughput is 34 kbps on 3rd observation. Based on Table 2
it can be concluded that the outermost areas have Internet access speeds throughput by 73.9% in the first area. This shows that
the outer area Internet capacity is quite good.
In addition to measuring throughput in some areas, we also do the measurement values of latency on WiFi mesh network.
Latency is the time it takes the network to respond to requests for an entry device to the network. This measurement is done by
accessing the website with a mobile phone in the outermost regions and then the server will measure the percentage of network
response time and data loss. Measurements lasted for 4 minutes with 10 times the number of observations and the number of
data packets varies accessed. Values latency measurement results can be seen in Table 3.
Table 3. Measurement of the value of latency in the outermost area
Examination Min
(ms)
Max
(ms)
Average
of delay
(ms)
Packet
loss
(%)
1 26 71 28 1
2 26 113 29 1
3 25 145 29 0
4 26 113 28 0
5 26 74 28 1
6 26 122 29 0
7 26 112 28 0
8 26 118 29 1
9 26 126 29 0
10 26 122 29 1
Average 28,6 0.5
Table 3 shows the average value of the outermost area latency is 28.6 ms. The maximum value of measured latency is 145 ms
on the 3rd observation. A minimum value of the measured latency is 25 ms on 3rd observation. The average percentage of data
loss during the measurement time is 0.5 percent. This shows that the mesh WiFi network that is built has a fast response time
with relatively small levels of data loss.
4. Conclusion
WiFi mesh infrastructure development in IoT application has been successfully carried out. The quality of the internet's
infrastructure has an average throughput value at 110.5 kbps in gateway area and 81.7 kbps in the outermost area are still able to
access the internet via WiFi mesh router. Moreover, the average latency measured in this outermost area of 28.6 ms with an
average percentage of data loss of 0.5 percent. Designing Client WiFi IoT also fared well. These nodes are small, low-power
and reliable. By connecting the WiFi client with electronic devices, meter-meter, and sensors, we have been able to observe and
control these devices through android or a computer server.

337
Rifki Muhendra et al. / Procedia Engineering 170 ( 2017 ) 332 – 337

References
[1] Lei Qin and Dongmei Zhao, Channel Time Allocations, and Handoff Management for Fair Throughput in Wireless Mesh Networks.IEEE Transactions on
Vehicular Technology, vol. 64 , no 1 (2015) pp.315 – 326
[2] Huang-Chen Lee and Hsiao-Hsien Lin, Design and Evaluation of an Open-Source Wireless Mesh Networking Module for Environmental Monitoring, IEEE
sensors journal, vol. 16, no. 7 (2016) pp.1-4
[3] Rifki Muhendra, Husein, Maman Budiman, and Khairurrijal, Development of digital water meter infrastructure using wireless sensor networks, Conference
on fundamental and applied science for advanced technology (AIP Conf. Proc. Yogyakarta, 2016) pp.1-6
[4] M. Karthikeyan, V. Saravanan, and S. Vijayakumar, Cloud based automatic street light monitoring system. International Conference on Green Computing
Communication and Electrical Engineering (ICGCCEE) (IEEE,India, 2014) pp.1-6
[5] Petros Spachos and Dimitrios Hatzinakos, Real-Time Indoor Carbon Dioxide Monitoring Through Cognitive Wireless Sensor Networks. IEEE sensors
journal, vol. 16, no. 2 (2016) pp. 506-514
[6] http://www.atmel.com/devices/atmega328.aspx
[7] https://nurdspace.nl/images/e/e0/ESP8266_Specifications_English.pdf
[8] https://github.com/mqtt/mqtt.github.io/wiki
[9] https://wiki.openwrt.org/toh/tp-link/tl-mr3020
[10] Axel Neumann, Ester Lopez, and Leandro Navarro. An evaluation of bmx6 for community wireless networks. In 1st International Workshop on
Community Networks and Bottom-up-Broadband (CNBuB, Barcelona, 2012), pp. 651–658