See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/347403060

Follow Me : A Human Following Robot using Wi-Fi Received Signal Strength

Indicator

Chapter · December 2020

DOI: 10.1007/978-981-15-8289-9_57

CITATIONS READS
10 2,356

5 authors, including:

Vasantha Geetha Sanket Salvi

National Institute of Technology Karnataka Dr. Vishwanath Karad MIT World Peace University
31 PUBLICATIONS 300 CITATIONS 33 PUBLICATIONS 242 CITATIONS

SEE PROFILE SEE PROFILE

Gurdeep Saini Naveen Yadav

National Institute of Technology Karnataka National Institute of Technology Karnataka
3 PUBLICATIONS 11 CITATIONS 4 PUBLICATIONS 17 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Sanket Salvi on 20 September 2022.

The user has requested enhancement of the downloaded file.

 Follow Me : A Human Following Robot using
Wi-Fi Received Signal Strength Indicator

Geetha V, Sanket Salvi, Gurdeep Saini, Naveen Yadav, and Rudra Pratap
Singh Tomar

Department of Information Technology,

National Institute of Technology, Surathkal, Karnataka
[email protected], [email protected],
[email protected], [email protected],
[email protected]

Abstract. Modernization has brought up significant changes in the way

humans used to communicate with devices which aid in reducing stren-
uous work. Robotics has made life simpler by automating some specific
tasks, one such application of robotics is explored in this paper. Human
Following Robot(HFR) can be used to carry heavy goods or simply fol-
low its user to act as an interface between IoT connected devices. In
this paper, a Low-Cost Prototype of Human Follower Robot is proposed
which uses Received Signal Strength Indicator of Wi-Fi to establish its
task.

Keywords: Wi-Fi signal · RSSI value · Serial communication · Human

Following Robot

1 INTRODUCTION
For developing countries like India, where manual labor is one of the major
employment option, the amount of money earned is comparatively very less
considering the amount of work carried out. This type of work majorly involves
manual loading of goods and transport. Similar, situation can also be observed
in military, where ammunition and food material have to be carried over longer
distances via difficult terrains. Thus, there is need of providing some additional
smart tools and mechanisms which could aid transportation without the need
for changing existing work structure. The growing industry of Internet of Things
(IoT) based devices can be considered as a possible solution which provides its
users with features such as remote control, monitoring and data analysis which
thereby enables automation of the tasks. Thus, introduction of fleet of load
carrying robots could significantly offload the arduous tasks of manual labour
and soldiers. To achieve this, while designing the system the technical aspects
such as optimum route selection, obstacle avoidance, etc, must be addressed.
Addressing technical issues also adds cost to the manufacturing of such system.
Thus, there is need of low cost alternatives to build robots that could accomplish
same task with lesser computation. One logical solution to this problem is to let
 Authors Suppressed Due to Excessive Length

human take the intelligent decision of route selection and obstacle avoidance
while the robot can simply follow the human.

2 Related Work

In this section, the work done by various authors in the domain of Human
Following Robot is highlighted.
In paper[1], a low cost, compact antenna for Wi-Fi is built to remotely control
a robot. This work has provided basic understanding of how the modern days Wi-
Fi modules work in principle. The necessary parameters need to be considered
for estimating directions in robots are explained well in this paper.
A vision based localization and navigation technique for robot is proposed in
paper[2] and a distributed system based mobile robot is proposed in paper[3].
Both these works highlight the major task in human following robot which is
to distribute the compute intensive tasks such as localization, path planning
and speech recognition so that the delay in the response of the robot can be
minimised as much as possible.
In paper[4][5], a SONAR based distance sensing approach is used for design-
ing a Human Following Robot. The proposed approach continuously finds the
distance between the sensor and human and tracks the change in the human
motion within its range. Thus, by moving the robot in the direction of change
in motion, the robot ends up following the human. However, this system fails
when multiple moving objects are present. To address this issue only one user
can be tagged with the robot, thus despite the presence of multiple moving ob-
jects/humans the system would work. However, it requires more resources and
computation intensive computer vision algorithm. Our proposed technique uses
the concept of single human tracking irrespective of multiple moving objects
with less compute intensive method.
In the paper[7], RFID is used for tracking the autonomous identity such
as human or robot within the indoor environment. The paper clearly show the
difference between locating the indoor identity using GPS and RFID, and results
shows that RFID worked well for close environment. In paper[8] proposed basic
path planning algorithm so that the robot can follow the human and can avoid
the obstruction by choosing the new path automatically. The robot is tested
in virtual robot experiment platform and results were satisfactory. These works
inspired the design of our proposed system.

3 Proposed System

In this section, the design and working of the proposed system is discussed. The
main motivation for our proposed work is to design a low cost HFR system.
Therefore, usage of existing easily available components for building HFR is a
major task as it directly effects on cost of the system. Compared to earlier work
done where the major power consumption depends on complex image processing
 Title Suppressed Due to Excessive Length

and extra hardware components such as additional sensors and Infrared cam-
era [4], our proposed work uses less computational power as power consumption
mainly depends only for calculation of left and right RSSI values and its com-
parison. Our proposed system uses mobile phone of user and two Wi-Fi modules
with onboard micro controller for decision making.

Fig. 1: Block Diagram of Proposed System

Fig.1 shows the basic block diagram of the proposed system. The mobile
phone of the person to be followed is connected to HFR via Wi-Fi. The HFR
subsystem has components such as left Wi-Fi module and right Wi-Fi module
which is connected to mobile phone at one end and micro-controller on the other
end. The micro-controller controls motor movement via motor driver.
It was observed that, the received signal strength of Wi-Fi module was
sensitive to the change in distance between module and its connected Wi-Fi
router/hotspot. By leveraging to this fact, it was decided to use RSSI value for
designing HFR. A pair Wi-Fi connected micro-controller can be used to achieve
this task. Once the pair of Wi-Fi module is connected to the desired user’s mo-
bile hotspot, the controller will decide upon the direction and distance between
the user based on the observed RSSI value. The concept is explained in detail
as shown in Fig.2
 Authors Suppressed Due to Excessive Length

Fig. 2: Basic working diagram of follow me robot

Fig.2 highlights the basic working flow of our proposed system. As the HFR
is powered, the Wi-Fi modules gets connected to a pre-defined Mobile Wi-Fi
Hotspot which is presumably of the user behind which the HFR has to follow.
These, Wi-Fi modules are placed over the HFR in such a way that both are
on the extremes of each side. Once the HFR Wi-Fi modules are connected, the
proposed system continuously calls for checking the RSSI value at respective
modules. These values change based on the distance between Hotspot hosting
Mobile Device and the HFR. The Microcontroller connected to HFR Wi-Fi mod-
ules continuously monitors this change and based on the coded logic it makes
appropriate decisions. The micro controller is also connected to a motor driver
which is used to control speed and direction of rotation of connected wheels.
Based on the programmed logic, the wheels will be rotated. Thus, if the RSSI
value of one side of HFR Wi-Fi module is more compared to other, it implies
that the Mobile phone is closer to one side, thus the other side motor will move
fast in order to balance the RSSI at both Wi-Fi modules. Also, if the RSSI value
at both HFR Wi-Fi modules is equal but less than the threshold, threshold is
 Title Suppressed Due to Excessive Length

used to measure the distance from the human if it is greater the limit then HFR
will stop then, the HFR will move forward to achieve the desired RSSI value.
Same principle is used when the mobile device gets too close to HFR to move it
backwards.

4 Requirement Analysis

Table.1 highlights the hardware and and software requirement for the implemen-
tation of our proposed system.

Table 1: Hardware and Software Requirements

Component
Usage
Name
An Opensource programmable board for Atmega microcontroller family.
Arduino
This The board provides set of analog and digital pins which can be
UNO
connected to take input from or give output to other devices.
It is esp8266 family programmable board with onboard Wi-Fi capability.
NodeMCU We have used two nodemcu to calculate the RSSI value at the right side
esp8266 and left side and after calculating these values, node mcu send it to the
arduino using serial communication.
L298D is a 16 -pin Integrated Circuit allows the DC motor to drive in
L298D both the direction according to our requirement. It handles the movement
Motor of the 2 DC motors at a time irrespective of the direction.Arduino supply
Driver the signal to the L298D then it send the power to the DC motor
according to signal
12 volt DC Motor are used to rotate the wheels of the robot car whose
DC Motor speed is 300 RPM and rated torque value is 1.2 Kg-cm, four such type
of motors are used to drive the HFR
Arduino IDE A programming tool for Arduino and NodeMCU

5 Implementation

The section highlights the implementation of proposed system along with the
detailed discussion on circuit diagram, connection details of components and
algorithmic implementation of the system. Fig.3, shows the connections of overall
components involved in building proposed system. Here, the NodeMCUs are
connected with Arduino UNO via software serial ports. This enables Arduino
to read the RSSI values printed at serial ports of each NodeMCU. A L298D
motor driver is externally powered by 12v battery and it is controlled by digital
PWM pins of Arduino. This allows to control the power supplied from battery
to motors and thus control speed of HFR. Arduino can vary the direction of
rotation of these wheels. NodeMCUs and Arduino are also externally powered
by battery.
 Authors Suppressed Due to Excessive Length

Fig. 3: Circuit Diagram of Human Following Robot

Algorithm 1 Algorithm used in Arduino for Direction Detection

1: countL = RSSI value received from left nodemcu
2: countR = RSSI value received from right nodemcu
3: if(countL<countR)
4: then rotate right()
5: if(countL>countR)
6: then rotate left()
7: if(countL == countR and countL≤ threshold)
8: then move forward();
9: if(countR≥threshold and countL≥threshold )
10: then stop();

In Algorithm 1, countR is RSSI value detected by right Nodemcu and countL

is RSSI value detected by left Nodemcu. The motion of the human is detected
based on RSSI values and by comparing the same with threshold value. If both
the RSSI values are greater than certain threshold value, then HFR will stop.
High RSSI value indicate that human is very near to robot. This algorithm plays
major role in the HFR as it performs the decisions of HFR such as left turn,
right turn, forward move and backward move.

In algorithm 2, our objective is to send the RSSI value from Nodemcu to Ar-
duino. As the algorithm is repeated at left and right side Nodemcu, explanation
of any one will be sufficient for understanding. From the left nodemcu, we get
the RSSI value. To send RSSI value to Arduino, a JSON object called as rootL
is created, which takes value from the serial monitor and stores in the object
by using it’s countL attribute. This object is then sent to Arduino via Software
Serial communication.
 Title Suppressed Due to Excessive Length

Algorithm 2 Algorithm used in left nodemcu

1: connect to wifi using password and signal name
2: countL = WiFi.RSSI();
3: StaticJsonBuffer<1000> jsonBuffer;
4: JsonObject& rootL=jsonBuffer.createObject();
5: rootL[”countL”] = countL;
6: send this value to arduino using serial communication

6 Results and Discussion

Table 2: RSSI value detected

RSSI value observed
distance-28cm R -15cm R 0 15cm L 28cm L
in cm
15cm LN=-44, LN=-43, LN=-34, LN=-35, LN=-39,
RN=-31 RN=-28 RN=-34 RN=-41 RN=-48
30cm LN=-44, LN=-41, LN=-39, LN=-51, LN=-50,
RN=-35 RN=-23 RN=-40 RN=-48 RN=-44
45cm LN=-44, LN=-41, LN=-39, LN=-51, LN=-50,
RN=-35 RN=-23 RN=-40 RN=-48 RN=-44
60cm LN=-44, LN=-44, LN=-42, LN=-57, LN=-43,
RN=-40 RN=-42 RN=-42 RN=-53 RN=-42
70cm LN=-68, LN=-50, LN=-46, LN=-60, LN=-51,
RN=-55 RN=-52 RN=-44 RN=-52 RN=-47

The performance of the proposed system is tested using a manually marked

surface by constructing different region of 15cm square. The RSSI values at both
side of HFR is noted for the manually marked surface region. Table. 2 shows the
observed values in each of the manually marked regions. HFR is placed at origin
and Mobile Phone is kept in different regions to observe the RSSI values. In
table 2, L and R denotes the left and right side of the HFR respectively. The
first column denote the forward straight distance in centimeter from the robot
and we are measuring the RSSI value after the 15cm distance gap. We observed
that as the distance increases, the RSSI value decreasing generally. But we have
also observed that in some cases where the row and column of the table 2 are
(2,4),(2,5),(3,4),(3,5),(4,4),(4,5),(5,4),and (5,5), the values are getting opposite
which means that left RSSI should be greater than the right RSSI. It is opposite
because of increased distance of human and and weak Wi-Fi signal strength.
For example, lets take (5,4)=(fifth row, fourth column) it means the box which
present at 70 cm straight and 15cm Left and at (5,4) values are LN=-60,RN=-52.
In the figure 4, the RSSI values of left node and same for right node with
respect to the displacement of the hot-spot devicep as describe in the table 2 is
plotted. The displacement is calculated as d = x2 + y 2 where d is the displace-
ment, x and y are the coordinates of hot-spot device considering the centre of
 Authors Suppressed Due to Excessive Length

Fig. 4: Bar graph for the RSSI values with respect to displacement of Hot-spot
device.

HFR as origin.The values of x and y can be obtained from the table 2. Consider
the coordinates (4, 1) the x coordinate is -28cm and y coordinate is 60cm. So,
the displacement calculated by the equation is 66.21cm and the clock is in the
left side of origin so we are considering it as negative and RSSI values of left
node is -44 and same for right node is -40. The right node is colored as orange
color and left node is colored as blue as mentioned in the graph as well. The
hardware components of the prototype design are as shown in figure 5.

Fig. 5: Side and Top View of HFR

7 Conclusion

The proposed HFR using the low cost components is implemented and its work-
ing is demonstrated successfully. Through experiment we have observed that
HFR successfully followed the human up to a significant mark. Proposed HFR
 Title Suppressed Due to Excessive Length

follows the direction in which human/mobile device is moving, which means the
HFR moves in the direction of hot-spot device. We have also observed that as
the distance of the mobile hot spot increases from the HFR. HFR fails to detect
the correct RSSI values on the left and right Nodemcu. This is majorly due to in-
creased interference and weak Wi-Fi signal strength. The limitation in proposed
system is that the weak strength of Wi-Fi signals results in smaller operational
distance. Hence, only RSSI based solution is not enough and the work can be
extended by using filters for more precise and stable values.

8 Acknowledgement
This is a pre-print of an article published in ICT Systems and Sustainability 2020
Vol. 1. The final authenticated version is available online at: https://doi.org/10.1007/978-
981-15-8289-9 57

References
1. Marcelo B. Perotoni , Beatriz E. Garibello, Silvio E. Barbin “ An IEEE802.11 Low
Cost Planner Antenna for a Mobile Robot”, 2006.
2. Gross, H.-M., Boehme, H.-J., and Wilhelm T., “Contribution to vision bas ed
localization, tracking and navigation methods for an interactive mobile service-
robot,” In the Proceedings of the IEEE International Conference on Systems, Man,
and Cybernetics, pp. 672–677, 2001.
3. Andrey A. Loukianov, Hidenori Kimura, Masanori Sugisaka. “Implementing dis-
tributed control system for intelligent mobile robot” 8th International Symposium
on Artificial Life and Robotics, Oita, Japan, January 24– 26 2003.
4. Quoc Khanh Dang, Young Soo Suh. “Human following robot using Infrared Cam-
era.” 2011 11th International Conference on Control, Automation and Systems,
BGyeonggi-do, South Korea, 2011.
5. Wei Peng, Jingchuan Wang, Weidong Chen, “Tracking Control of Human-
Following Robot with Sonar Sensors” Intelligent Autonomous Systems 14. IAS
2016. Advances in Intelligent Systems and Computing, vol 531. Springer, Cham.
6. Che-Wen Chen, Shih-Pang Tseng , Yao-Tsung Hsu , Jhing-Fa Wang “Design and
implementation of human following for separable omnidirectional mobile system of
smart home robot” 2017 International Conference on Orange Technologies (ICOT)
IEEE, Singapore, Singapore
7. Ricardo Tesoriero, José A. Gallud, Marı́a D. Lozano, Vı́ctor M. R. Penichet “Track-
ing Autonomous Entities using RFID Technology” IEEE Transactions on Con-
sumer Electronics, Vol. 55, No. 2, MAY 2009
8. ’Simulation Of A Human Following Robot With Object Avoidance Function’ 12th
South East Asian Technical University Consortium,2018. Shimoyama Mirai, Mut-
suhira Nobuto,
9. Suzuki Kaoru, ‘Human Characterization by a Following Robot Using a Depth
Sensor(I),” in IEEE/SICE International Symposium on System Integration, 2017.

View publication stats