2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)
DESIGN OF HUMAN BEHAVIOR AUTOMATIC LAMP SWITCH
WITH BLYNK PLATFORM
Muhammad Nasrun
Adiatma Manglolo Pasau
School of Electrical Engineering
School of Electrical Engineering
Telkom University
Telkom University
nasrun@telkomuniversity.ac.id
adiatmamp@student.telkomuniversit
y.ac.id
Abstract— The use of Internet of Things (IoT)
technology provides comfort and convenience for
technology users. In this research, an automatic light
switch device used the Light Dependent Resistor sensor
as an automation for light intensity and AI server
assistance for human habits. In this tool there are two
rules that govern the automation of the Decision Rule
and Priority Rule. From the testing of this tool, the two
rules have shown an accuracy of 100%.
Keywords— automatic light switches, blynk platform,
internet of things
I. INTRODUCTION
With the development of technology and the Internet
of Things makes it easier for technology users to carry
out any activities related to technology. The
development of the Internet of Things in every aspect
of life makes it easier for small things to be done. [1]
As in this Project, the making of a custom human light
switch uses the Blynk platform as an Inter of Things
and NodeMCU v1 broker as its microcontroller that
will be connected via an internet connection via Wi-Fi.
Lighting settings usually use an ordinary switch to turn
lights on or off by operating the switch manually.
Lighting a room will turn on or turn off automatically
following the habits carried out by the user of the room
so that the switch becomes automatic, assisted also by
the presence of Light Dependent Resistor sensors to
help the machine detect the intensity of light in the
room, so it can facilitate the control of the lights. Not
only using a machine in this room will be given a
switch button that is useful for turning off and turning
on the lights manually and also the use of servers to be
able to turn on and off remotely using a smartphone
application.
To make it easier for users to control the lights, this
tool can be a solution to facilitate the control of the
lights automatically and flexibly. This tool itself has
accuracy and response time that is efficient in
operation.
II. LITERATURE STUDY
Research on automatic lights, in this study using
Arduino mega as a microcontroller and Light
Dependent Resistor sensor as an automatic benchmark
of lights in measuring light intensity [2]. Automatic
light experiment but only as a night light, the use of
Light Dependent Resistor sensors is only to turn on
978-1-7281-4865-6/19/$31.00 ©2019 IEEE 150
Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
Casi Setianingsih
School of Electrical Engineering
Telkom University
Setiacasie@telkomuniversity.ac.id
the lights at night and also help the Passive Infrared
sensor to find out human existence [3].
Researching the effectiveness of Light Dependent
Resistor sensors for use in automatic lights. In the
study said that the Light Dependent Resistor sensor
has been able to replace old technology and is very
effective, therefore the researchers used the Light
Dependent Resistor sensor in this study [4]. Research
on automatic lights using Passive Infrared sensors, in
this study can be seen that the use of Passive Infrared
sensors is very useful if installed in a room that has
activity [5].
III. RELATED TECHNOLOGIES
A. Blynk
Blynk is a server service used to support the Internet
of Things project. This server service has a mobile
user environment both Android and iOS. Blynk
Application as a support for IoT can be downloaded
through Google play. Blynk supports a variety of
hardware that can be used for the Internet of Things
project. Blynk is a digital dashboard with a graphical
interface facility in making the project. Adding
components to Blynk Apps by Drag and Drop, making
it easier to add Input / output components without the
need for Android or iOS programming capabilities.
Blynk was created with the aim of controlling and
monitoring hardware remotely using internet or
intranet data communication (LAN network). The
ability to store data and display data visually using
either numbers, colors or graphics makes it easier in
making projects in the Internet of Things [6].
B. NodeMCU
NodeMCU is an open source IoT platform. Consisting
of hardware in the form of a System On Chip
ESP8266 from ESP8266 made by Espressif System,
also the firmware used, which uses the Lua scripting
programming language. The term NodeMCU by
default actually refers to the firmware used rather than
the hardware development kit [7].
C. Light Dependent Resistor
Light Sensor or Light Dependent Resistor is a resistor
that has a resistance value or resistance value that
depends on the intensity of the light it receives. When
it is clear the Light Dependent Resistor resistance
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

value will decrease and in the dark the resistance value The mobile application is connected as a visual for
will rise. In other words, the Light Dependent Resistor lamp status and as a remote input using the Blynk
functions to deliver an electric current if it receives a mobile application. Then the house lights that connect
certain amount of light intensity (Bright Conditions) through the Relay.
and inhibits electric current in dark conditions. The
received light affects the rise and fall of the resistance A. Hardware Design
value [8]. In this study, researchers used two rooms as
experiments, so a miniature room prototype was made,
D. Forward Chaining each room measuring 30 cm. In each - each room there
Forward Chaining is one of the expert system methods, is 1 home lamp, Light Dependent Resistor sensor, and
Forward Chaining is itself a search method or a a switch that is connected to the microcontroller
forward tracking technique that starts with existing installed in a miniature room. From each room the door
information and combines rules that produce is given to get the light intensity from each room. To
conclusions. In Forward Chaining must use the data drain electricity the appliance is fitted with an electric
approach first to describe the conclusions. In this socket branch to turn on the microcontroller and the
approach starts from the available information or basic house lights. Each component - a component that is
ideas [9]. connected directly to the microcontroller using a
jumper cable. Example in Fig.1.
E. Rule Based System
Rule Based System is a system that uses rules to show
their knowledge from rules. This theory starts with a
simple technique starting with a basic rule that
contains all the knowledge of the problems faced
which are then encoded into if-then rules that contain
data, statements and initial information [10].
IV. OVERVIEW
The general description for this automatic light switch
explains how this tool works, this tool sends data, a
series of these tools, the design of mobile applications,
learning modes and modes of human habits and the
rules used in this tool. Fig. 2. Hardware Design

B. Blynk Aplication Design

In the mobile application there are two tabs where
each tab represents one lamp, for this study using two
tabs because it uses two lights. Each tab consists of a
led indicator as a visual on a home light for mobile
applications and a button switch used for remote home
light controllers.

Fig. 1. System Overview

Broadly speaking, in this tool the microcontroller is
connected to Capacitive Touch where the Capacitive
Touch works as a physical switch on the device. The
microcontroller is also connected to the Light Fig. 3. Blynk Application Design
Dependent Resistor Sensor and Relay, where the Light
Dependent Resistor Sensor as the decision maker to C. Store Data
determine the light or darkness of a room and the
Data transmission between Automatic Light Switches
Relay as a breaker or current connector for the light
with Human Habit Server where in this condition the
switch.
device sends lamp status data to become learning on
The tool is also connected to the server to send
the server and vice versa the tool receives input from
learning data or receive input from the server. To
the server when it has turned into a mode of human
connect devices and servers using an internet network
via wifi by a microcontroller.

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 151

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

habits. Data from Device is sent through antares microcontrollers, switches, Light Dependent Resistor
server. Figure 6. sensors and Relays. In the picture above shows the
part in the prototype room that is used automatic
lights, where in this section there are two lights on the
top that can be open or closed, and the room divider.
E. Device Mode’s
There are two modes on this device, the first is
learning mode and the second is human behavior
mode. for the process, the device will read the status
of the AI Server to determine which device is in a
certain mode. for more details can be seen on Figure
Fig. 4. Store Data Program 8.
In Figure 5 can be seen the data received on the
antares server.

Fig. 5. Antares Server

Antares is a local server in Indonesia as a platform for Fig. 8. Flowchart Mode in this device.
supporting the internet of things and functions as a 1) Learning Mode
MQTT Broker. Learning mode, this is the condition when the device
only gets input from Switches and Sensors. This tool
is used as input for servers in learning human
behavior.

Fig. 6. Store Data

In the figure 6 above shows that the device will send
data to the antares server to save human activity data,
then from the antares server it will be the learning data
for the AI server. the results of the AI server will be
sent through Blynk as an input for AI.
D. Implementation
In designing this prototype using a duplex board
designed like a two-room model, and each of the
rooms has a gap like a door to see the condition of the
lights and lighting in the room.

Fig. 9. Learning Mode

In Figure 9. we can know that the tool will read the

input from the switch then it will be diverted to the
Fig. 7. Prototype Of Implementation Device mobile application if there is no input from the switch,
In Figure 7. shows the top of the prototype room that then if the mobile application does not have an input it
is used for automatic lights, and there are

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 152

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

will be switched to the Light Dependent Resistor The Table 1 above shows the decision-making rule in
sensor. the learning mode that will be used as a determinant of
the input received by the device and this rule will
2) Human Behavior Mode determine the condition of the next lamp. In this table
Human Behavior Mode, this is the condition when the in the switch column, mobile and lamp state the status
server has learned human habits so that the tool gets of each component. Then on the row in blue indicates
additional input provided by the server's human habits. the status change, for example OFF → ON, it states
that there is a change in status from off to on. In the
column results declare that the results after the status
changes occur. For example on the status of the switch
is off, the mobile is off, the sensor is dark and the
lamp is off. Then after changing the status of the
switch to on (OFF→ ON) it will change the status of
lamp and mobile.
Table 2. Decision Maker Rule in Human Behavior
Mode
AI Switch Mobile Sensor Lamp Result
ON → Only change a AI
OFF OFF DARK OFF
OFF stat
OFF → Lamp, Switch,
OFF OFF DARK OFF
ON Mobile = ON
ON → Only change a AI
OFF OFF LIGHT OFF
OFF stat
OFF → Lamp, Switch,
OFF OFF LIGHT OFF
ON Mobile = ON
ON → Lamp, Switch,
ON ON DARK ON
OFF Mobile = OFF
OFF → Only change a AI
ON ON DARK ON
ON stat
ON → Lamp, Switch,
ON ON LIGHT ON
OFF Mobile = OFF
OFF → Only change a AI
ON ON LIGHT ON
Fig. 10. Human Behavior Mode ON stat
OFF →
In Figure 10. we can know that the tool will read ON
ON
OFF DARK OFF Lamp, Mobile = ON
starting from the AI server and then no input will be OFF →
ON OFF LIGHT OFF Lamp, Mobile = ON
switched on the input from the switch then it will be ON
OFF →
transferred to the mobile application if there is no OFF
ON
OFF DARK OFF Lamp, Mobile = ON
input from the switch, then if the mobile application OFF
OFF →
OFF LIGHT OFF Lamp, Mobile = ON
ON
does not have input it will be switched to the Light ON → Lamp, Mobile =
Dependent Resistor sensor . ON ON DARK ON
OFF OFF
ON → Lamp, Mobile =
ON ON LIGHT ON
F. Decesion Maker Rule OFF OFF
ON → Lamp, Mobile =
In this rule using the Rule Base System, which has OFF
OFF
ON DARK ON
OFF
been determined by the researcher rule, there are 2 OFF
ON →
ON LIGHT ON
Lamp, Mobile =
OFF OFF
parts, namely the Rule, Decision Making when the OFF →
ON OFF DARK OFF Lamp, Switch = ON
mode of Learning and Habits is Human. Here's the ON
rule : OFF →
ON OFF LIGHT OFF Lamp, Switch = ON
ON
OFF →
OFF OFF DARK OFF Lamp, Switch = ON
Table 1. Decesion Maker Rule In Learning Mode ON
NO Switch Mobile Sensor Lamp Result OFF →
OFF OFF LIGHT OFF Lamp, Switch = ON
OFF → Lamp = ON , Mobile = ON
1 OFF DARK OFF ON →
ON ON ON ON DARK ON Lamp, Switch = OFF
OFF → Lamp = ON , Mobile = OFF
2 OFF LIGHT OFF ON →
ON ON ON ON LIGHT ON Lamp, Switch = OFF
ON → Lamp = OFF , Mobile = OFF
3 ON DARK ON ON →
OFF OFF OFF ON DARK ON Lamp, Switch = OFF
ON → Lamp = OFF , Mobile = OFF
4 ON LIGHT ON ON →
OFF OFF OFF ON LIGHT ON Lamp, Switch = OFF
OFF → Lamp = ON , Switch = OFF
5 OFF DARK OFF LIGHT → Lamp, Switch,
ON ON OFF OFF OFF OFF
OFF → Lamp = ON , Switch = DARK Mobile = ON
6 OFF LIGHT OFF DARK → Only change a
ON ON OFF OFF OFF OFF
ON → Lamp = OFF , Switch = LIGHT sensor stat
7 ON DARK ON LIGHT → Lamp, Switch,
OFF OFF ON OFF OFF OFF
ON → Lamp = OFF , Switch = DARK Mobile = ON
8 ON LIGHT ON DARK → Only change a
OFF OFF ON OFF OFF OFF
DARK → CHANGE SENSOR LIGHT sensor stat
9 OFF OFF OFF LIGHT → Only change a
LIGHT STATUS OFF ON ON ON
LIGHT → Lamp , Switch, Mobile = DARK sensor stat
10 OFF OFF OFF DARK → Lamp, Switch,
DARK ON OFF ON ON ON
DARK → Lamp , Switch, Mobile = LIGHT Mobile = OFF
11 ON ON ON LIGHT → Only change a
LIGHT OFF ON ON ON ON
LIGHT → CHANGE SENSOR DARK sensor stat
12 ON ON ON ON ON ON DARK → ON Lamp, Switch,
DARK STATUS

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 153

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

AI Switch Mobile Sensor Lamp Result Variable Variable

Indicator Rule
LIGHT Mobile = OFF Indicator Rule
The Table 2 above shows the decision-making rule in Switch On E IF K AND I THEN V R5
Switch OFF F IF K AND J THEN U R6
the Human Behavior mode that will be used as a Switch Trigger G IF O AND M THEN X R7
determinant of the input received by the device and Switch Idle H IF O AND N THEN W R8
IF C AND G AND K AND O
this rule will determine the condition of the next lamp. Mobile On I
THEN Y
R9
In this table it is also the same as Table I, only Mobile Off J IF C AND K AND O THEN Y R10
Mobile
increasing the status of the AI Server. Trigger
K IF C AND G AND O THEN Y R11
Mobile Idle L IF C AND G AND K THEN Y R12
G. Priority Rule Sensor Dark M IF C AND O THEN Y R13
Sensor Light N IF C AND G THEN Y R14
In the Priority Rule, using Forward Chaining on this Sensor Trigger O IF C AND K THEN Y R15
rule there are also 2 rules, namely the rule for learning Sensor Idle P IF G AND K AND O THEN Z R16
modes and human habits. AI Lamp On Q IF G AND K THEN Z R17
AI Lamp On R IF G AND O THEN Z R18
Table 3. Priority Rule in Learning Mode Switch Lamp
S IF K AND O THEN AA R19
Variable Variable Off
Indicator Rule Switch Lamp
Indicator Rule T IF D AND H AND L THEN AB R20
On
Switch On A IF C AND A THEN N R1 Mobile Lamp
Switch OFF B IF C AND B THEN M R2 U IF Y AND R THEN AC R21
Off
Switch Trigger C IF G AND E THEN P R3 Mobile Lamp
V IF Y AND Q THEN AD R22
Switch Idle D IF G AND F THEN O R4 On
Mobile On E IF K AND I THEN R R5 Sensor Lamp
W IF Z AND T AND D THEN AC R23
Mobile Off F IF K AND J THEN Q R6 Off
IF C AND G AND K THEN Sensor Lamp
X IF Z AND S AND D THEN AD R24
Mobile Trigger G S R7 On
Mobile Idle H IF C AND G THEN S R8 IF AA AND V AND D AND H
AI Prio Y R25
Sensor Dark I IF C AND K THEN S R9 THEN AC
IF AA AND U AND D AND H
IF G AND K AND D Switch Prio Z R26
THEN AD
Sensor Light J THEN T R10
Mobile Prio AA IF AB AND X THEN AC R27
Sensor Trigger K IF D AND H THEN U R11 Sensor PrIo AB IF AB AND W THEN AD R28
Sensor Idle L IF S AND N THEN V R12 Lamp On AC
Switch Lamp Off M IF S AND M THEN W R13 Lamp Off AD
Switch Lamp On N IF T AND P THEN V R14
Mobile Lamp
Off O IF T AND O THEN W R15 In Table 4 the priority rule in human behavior mode
Mobile Lamp On P IF U AND R THEN V R16 can be seen as a rule for determining the priority of
Sensor Lamp Off Q IF U AND Q THEN W R17
Sensor Lamp On R each input.
Switch Prio S
Mobile Prio T
Sensor PrIo U
Lamp On V
Lamp Off W
In Table 3 the priority rule in learning mode can be
seen as a rule for determining the priority of each
input.

Fig. 11. Decision Tree of Priority Rule in Learning Mode

The Figure 11 shows the decision tree from the Fig. 12. Decision Tree of Priority Rule in Human Behavior Mode
priority rule of learning mode in Table III. And can be The Figure 12 shows the decision tree from the human
seen from this decision tree can provide conclusions behavior rule of learning mode in Table IV. And can
between lights off or not. be seen from this decision tree can provide
Table 4. Priority Rule in Human Behavior Mode conclusions between lights off or not.
Variable Variable
Indicator Rule
Indicator Rule
AI On A IF C AND A THEN R R1
AI Off B IF C AND B THEN Q R2
AI Trigger C IF G AND E THEN T R3
AI Idle D IF G AND F THEN S R4

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 154

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

V. EXPERIMENT AI Switch Mobile Sensor Lamp Result

Experiment
Result
OFF Mobile = OFF
A. Decision Maker Rule Experience OFF
ON →
ON LIGHT ON
Lamp,
Correct
OFF Mobile = OFF
Testing this rule is done each - each rule as much as 5 OFF → Lamp, Switch
ON OFF DARK OFF Correct
tests, and the results obtained are as follows: ON = ON
OFF → Lamp, Switch
Table 5. Decision Maker Rule Experiment in Learning ON OFF
ON
LIGHT OFF
= ON
Correct
Mode OFF OFF
OFF →
DARK OFF
Lamp, Switch
Correct
Experiment ON = ON
Switch Mobile Sensor Lamp Result
Result OFF → Lamp, Switch
OFF OFF LIGHT OFF Correct
OFF → Lamp = ON , ON = ON
OFF DARK OFF Correct
ON Mobile = ON ON → Lamp, Switch
ON ON DARK ON Correct
OFF → Lamp = ON , OFF = OFF
OFF LIGHT OFF Correct
ON Mobile = ON ON → Lamp, Switch
ON ON LIGHT ON Correct
ON → Lamp = OFF , OFF = OFF
ON DARK ON Correct
OFF Mobile = OFF ON → Lamp, Switch
OFF ON DARK ON Correct
ON → Lamp = OFF , OFF = OFF
ON LIGHT ON Correct
OFF Mobile = OFF ON → Lamp, Switch
OFF ON LIGHT ON Correct
OFF → Lamp = ON , OFF = OFF
OFF DARK OFF Correct
ON Switch = ON LIGHT Lamp,
OFF → Lamp = ON , OFF OFF OFF → OFF Switch, Correct
OFF LIGHT OFF Correct
ON Switch = ON DARK Mobile = ON
ON → Lamp = OFF , DARK
ON DARK ON Correct Only change
OFF Switch = OFF OFF OFF OFF → OFF Correct
a sensor stat
ON → Lamp = OFF , LIGHT
ON LIGHT ON Correct
OFF Switch = OFF LIGHT Lamp,
DARK CHANGE A ON OFF OFF → OFF Switch, Correct
OFF OFF → OFF SENSOR Correct DARK Mobile = ON
LIGHT STATUS DARK
Only change
LIGHT ON OFF OFF → OFF Correct
Lamp, Switch, a sensor stat
OFF OFF → OFF Correct LIGHT
Mobile = ON
DARK LIGHT
Only change
DARK OFF ON ON → ON Correct
Lamp , Switch, a sensor stat
ON ON → ON Correct DARK
Mobile = OFF
LIGHT DARK Lamp,
LIGHT CHANGE A OFF ON ON → ON Switch, Correct
ON ON → ON SENSOR Correct LIGHT Mobile = OFF
DARK STATUS LIGHT
Only change
ON ON ON → ON Correct
In Table above shows the test results from learning DARK
a sensor stat
mode. From the expected results in both modes, all the DARK Lamp,
rules have been fulfilled, on this rule 100% have been ON ON ON → ON Switch, Correct
LIGHT Mobile = OFF
successful.
In Table above shows the test results from both
Table 6. Decision Maker Rule Experiment in Human
modes, learning modes and Human Behavior mode.
Behavior Mode
Experiment
From the expected results in both modes, all the rules
AI Switch Mobile Sensor Lamp Result
Result have been fulfilled, on this rule 100% have been
ON
→ OFF OFF DARK OFF
Only change
Correct
successful.
a AI stat
OFF
OFF Lamp,
B. Priority Rule Experiment
→ OFF OFF DARK OFF Switch, Correct In this test the samples taken were tested 5 times from
ON Mobile = ON
ON each sample and the following are the results
Only change
→ OFF OFF LIGHT OFF
a AI stat
Correct obtained:
OFF
OFF Lamp, Table 7. Priority Rule Experiment in Learning Mode
→ OFF OFF LIGHT OFF Switch, Correct Input Experiment
Result
ON Mobile = ON Switch Mobile Sensor Result
ON Lamp, ON OFF DARK LAMP ON CORRECT
→ ON ON DARK ON Switch, Correct
OFF OFF IDLE LAMP OFF CORRECT
OFF Mobile = OFF
OFF
OFF IDLE LIGHT LAMP OFF CORRECT
Only change IDLE ON LIGHT LAMP ON CORRECT
→ ON ON DARK ON Correct
a AI stat
ON
ON Lamp,
→ ON ON LIGHT ON Switch, Correct The Table above shows the test results from learning
OFF Mobile = OFF mode. From the expected results in both modes, all the
OFF
→ ON ON LIGHT ON
Only change
Correct
rules have been fulfilled, on this rule 100% have been
a AI stat
ON successful.
ON
OFF
OFF DARK OFF
Lamp,
Correct Table 8. Priority Rule Experiment in Human Behavior
→ ON Mobile = ON
OFF Lamp, Mode
ON OFF LIGHT OFF Correct
→ ON Mobile = ON Input Experiment
Result
OFF
OFF
OFF DARK OFF
Lamp,
Correct AI Switch Mobile Sensor Result
→ ON Mobile = ON OFF ON ON DARK LAMP ON CORRECT
OFF Lamp,
OFF OFF LIGHT OFF Correct ON ON ON IDLE LAMP ON CORRECT
→ ON Mobile = ON
ON → Lamp, ON ON IDLE LIGHT LAMP ON CORRECT
ON ON DARK ON Correct OFF IDLE ON LIGHT LAMP OFF CORRECT
OFF Mobile = OFF
ON → Lamp, OFF ON IDLE IDLE LAMP OFF CORRECT
ON ON LIGHT ON Correct
OFF Mobile = OFF ON IDLE ON IDLE LAMP ON CORRECT
OFF ON → ON DARK ON Lamp, Correct

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 155

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.
2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

Input Experiment 2. In the Priority Rule each input on the

Result
AI Switch Mobile Sensor Result
Automatic Switch Light tool in the Learning
ON IDLE IDLE LIGHT LAMP ON CORRECT
Mode and Custom Mode is appropriate and
The Table above shows the test results from human running well. So if when two or more inputs
behavior mode. From the expected results in both enter the device there will be no crash.
modes, all the rules have been fulfilled, on this rule 3. For the response time of this Automatic Light
100% have been successful. Device, the average is only about 0.2
milliseconds from each input tested.
C. Response Time Experiment 4. This tool has fulfilled the purpose of this
In the Response Time testing scenario, each input of research, which functions in direct, remote and
each lamp will be tested five times and will be even automatic light control.
calculated on average of the total data per each input
and per each lamp.
REFERENCES
[1] M. Bounhas, H. Prad, M. Serrurier and K.
Mellouli, "A Possibilistic Rule-Based Classifier,"
IPMU 2012, 2012.
[2] A. A. Dahoud and M. Fezari , "NodeMCU V3
For Fast IoT Application Development," 2018.
[3] W. V T and Y. D. Pramudita, "Automatic Lamp
and Fan Control Based on Microcontroller,"
Journal of Physics: Conference Series, 2018.
Fig. 13. Response Time
[4] V. Sharma and R. Tiwari, "A review paper on
From the graph, we can find out that the average “IOT” & It‟s Smart Applications," International
response time on automatic lighting tools is below 0.4 Journal of Science, Engineering and Technology
seconds, except for sensors in light 2 because in 5 Research (IJSETR), vol. 5, 2016.
tests the sensors in lamp 2 had a long delay which [5] A. Rupnawar, "Study on Forward Chaining and
resulted in a significant surge of graphs up to 1.13 Reverse Chaining in Expert System,"
second. International Journal of Advanced Engineering
Research and Science (IJAERS), vol. 3, no. 12,
Response Time 2016.
[6] D. K. Rath , "Arduino Based: Smart Light
200
Control System," International Journal of
150 Engineering Research and General Science, vol.
Time (ms)

4, no. 2, 2016.
100 [7] K. R.A , S. Sriram K and P. D, "Automatic
Lighting Using Arduino and Passive Infrared
50
Sensor," International Journal of Computer
0 Science Trends and Technology (IJCST), vol. 6,
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 no. 5, 2016.
Push [8] Z. Mumtaz, S. Ullah, . Z. Ilyas, S. Liu, N. Aslam
and J. A. Meo, "Automatic Streetlights that Glow
on Detecting Night and Object using Arduino,"
100 ms 500 ms 1000 ms 2018.
Fig. 14. Multi Input [9] A. Jalan, . G. Hoge, S. Banaitkar and S. Adam,
From the test results above, we can conclude that the "Campus Automation Using ARDUINO,"
average of all Multi Input Response Time tests is International Journal of Advanced Research in
86.63 milliseconds with the fastest value of 60 Electrical, Electronics and Instrumentation
milliseconds and 169 millisecond late. Engineering, vol. 6, no. 6, 2017.
CONCLUSION [10] H. S. DOSHI and M. S. SHAH, "INTERNET of
THINGS (IoT): INTEGRATION of BLYNK for
1. Decision Maker Rule in Learning Mode runs DOMESTIC USABILITY," VJER-Vishwakarma
already in accordance with the expected output. Journal of Engineering Research , vol. 1, no. 4,
Likewise in the Human Behavior Mode, all the 2017.
tests are running according to the rules that
apply.

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 156

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 02,2020 at 16:35:01 UTC from IEEE Xplore. Restrictions apply.