International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056

Volume: 04 Issue: 04 | Apr -2017 www.irjet.net p-ISSN: 2395-0072

Human Follower Robot Using Kinect

Purvi Agarwal1, Pranjal Gautam2, Anmoal Agarwal3, Vijai Singh4
1Department of Computer Science and Engineering, IMS Engineering College, Ghaziabad, India
2Department of Computer Science and Engineering, IMS Engineering College, Ghaziabad, India
3Department of Computer Science and Engineering, IMS Engineering College, Ghaziabad, India
4 Professor, Department of Computer Science and Engineering, IMS Engineering College, Ghaziabad, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This paper throws light on the aspects of human the full body of the human standing before it by showing 20
following robot using Kinect technology. The goal of this study joints of the human body. The data calculated in the form of
is to eliminate the limitations of human robots by presentation the distance between the human and the Kinect camera is
of a Human follower robot which moves and is directly fed to the computer that is the bridge between the two
controlled by the human in front of the camera. This paper is systems- the robot and the Kinect.
focusing on the fact, how the Kinect sensor captures the 3-d
information of a scene and recognizes the human body by
retrieving the depth information. Microsoft Kinect is one of the
latest advancement in the field of human and computer
interaction (HCI). Various studies are being conducted using
Kinect in the field of computer vision and action recognition
techniques. In this process, the distance of human from the
Kinect camera is being calculated in the form of the depth
data, which is being Instantiated to move the robot forward
and backwards. The robot being used here is a two legged
robot powered by Arduino Uno microcontroller and a
Bluetooth module ( HC-05) for the interfacing and data
transfer unto the robot. The whole process has a wide range of
applications based upon application areas. The robot is
independent of the particular human standing in front of him
and will follow any other user after losing the field of view
during its movements.
Key Words: Human Follower Robot, Kinect Sensor,
Human Computer Interaction

1.INTRODUCTION Figure 1 : Kinect Body Joints

Robotics is a field that has seen much advancement in the 2. ROBOT PLATFORM
past few years and before. This field has found many
applications in the field of military, industry, medicine and
The mobile robot used here is a two wheeled robot
many others. The absence of robotics in daily use has led to having approximate height of 0.15cms which is powered by
the need of an intelligent robot. Thus the human following an arduino uno microprocessor. There are two 9 volt
robots come into action due to their capability as working batteries being used to power the whole system. The wheels
assistants for the humans and the further developments can are in direct connection to a dc motor for movement facility
see these carrying heavy loads in the various hauling forward and backward.[6]
activities.
Currently this field is under many researches which use Hc-05 bluetooth module is being used in the form of
various technologies so as to locate the person standing in wireless technology to make the system mobile, which can be
further improved by using infrared and laser communication
front of the camera using depth sensors and digital image
systems to increase the range which is currently 10 metres.
processing. And Kinect camera has a quite less area
limitation compared to the others along with its various The robot is capable of forward and backward motion
qualities, this makes it the best in the business.[7] currently obeying the movement of the human before it.
Kinect sensors are versed with human tracking capabilities,
which provide a full 3-D motion capture thus making it the
backbone of the human following systems. The accuracy of
the Kinect sensor can be judged by its capability to represent

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1635
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 04 | Apr -2017 www.irjet.net p-ISSN: 2395-0072

3. KINECT CAMERA

The Microsoft Kinect camera being used here in the

project is one of the recent technologies developed in the
field of image capturing and computer vision due to its 3d
image and scene capture, depth calculation and action
recognition facilities. The innovation behind the Kinect is a
mixture of hardware and software contained within the
Kinect sensor accessory that has various applications.[2]

Kinect has three basic components

1. Color vga video camera- This camera finds its

applications in the field of facial recognition by
Figure 3: Depth Recognition
2. detection of the three color components- red, green
and blue and hence its name – RGB camera. It stores
triple channel data in the resolution of 1280* 960.
4. METHODOLOGY
3. Depth sensor: The depth sensor present in this
works in any conditions to see anything regardless
of the light in 3-d[8] Human following: It is broken down into two steps:
4. Multi array microphone: The multi array
microphone present in the Kinect camera has 1. Distance calculation: Every point in the space has 3 values
function of isolating the different voices from (x,y,z) projected in the Cartesian coordinate system. The
various sources thus making it capable of voice point (0,0,0) is the position of the sensor. The measurements
recognition. The main advantage being it can of other points is measured with respect to the sensor only.
capture voices from a distance even. The depth is calculated by the z axis distance which is
5. Skeleton tracking: The Kinect skeleton tracking originally the distance between the point and the plane of
identifies the 20 joins present in the human body. the sensor. Thus the distance of the human standing in front
of the kinect is the z-axis distance.[1]

Bringing it all together

var point= body.Joints[JointType.Spine].Position

var distance= Length(point)[4]

Figure 2 : Kinect Camera

3.1 KINECT SENSOR

The range of this particular sensor is between 500mm to

nearly 4000mm. Infrared is used so that the kinect sensors
can see through glass. The infrared camera used is a
monochrome complementary metal oxide semiconductor
(CMOS) sensor. The main use is the quality of reconstruction Figure 4: Kinect Coordinate Mapping
of even a point in 3d.[3]
It produces a depth image map for the Infrared image. The
depth value consists of gray functional values and darker the
pixel means that the point is closer to the camera which
fades in color accordingly to the distance.

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1636
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 04 | Apr -2017 www.irjet.net p-ISSN: 2395-0072

5. FUTURE SCOPE

Construction of the human follower robots is just in its

starting phase. In future these can be seen as the ones
carrying loads for the people of several disciples.
Incorporation of large number of individuals will give a new
direction to the scope of human follower robots.

6. CONCLUSION

We would like to conclude the topic by saying that a lot of

further improvisations and improvements can be done so as
to increase the efficiency and the overall working of the
project. This system has a lot of applications throughout in
Figure 5 : Kinect Depth Calculation every field and our proposed methodology makes an
emphasis on control of the robot in human way thus making
2. Data transmission to robot: the task simpler. When the use of joysticks and wires are
exempted, it increases the in site qualities of the robot.
Using the depth value taken from the previous component
and the location of the human target from the human target ACKNOWLEDGEMENT
detection component(kinect), the robot is able to follow the
human target and at the same time check if the distance of We are thankful to each and every person involved directly
the human target and the thresholds value. The robot or indirectly in the completion of the project. We would like
classifies the movement in the following way[5]: to thank our mentor for devoting his time for our work.
Further the almighty for providing us with the strength and
If depth is between the range of 2-4 meters: - the power which guided us till the end.
 Key A is pressed on the keyboard automatically
without human touching.
 Robot automatically reads the key pressed and
REFERENCES
moves forward.
If depth is in the range of 1.5-2 meters: -
[1] Buniyamin, N. Wan Ngah, W.A Sarif,N.& Mohammad Z
 Key B is pressed automatically on the keyboard. (2011) -A Simple Path Planning Algorithm For
 Robot reads the value and stops. Autonomous Robot. International Journal Of Systems
If depth is in the range of 0.5-1.5 meters: Applications, Engineering And Development Issue 2,
 Key C is pressed on the keyboard. Volume 5
 Robot reads the value and moves backward. [2] Open NI User Guide. (N.D) Retrieved From Kinect Hacks:
http://www.kinect-hacks.com
[3] Pakinis , T. (2010, JUNE 29) Kinect Technical
Specifications Revealed. Retrieved from Computer And
Video Games.
[4] http://channel9.msdn.com/coding4fun/kinect/The-right-
way-to-measure-distances-with-the-kinect
[5] Alfred George Ang - Human Following Robot Using
Kinect Sensor- Presented at Research Congress 2013.
[6] Morioka. K .Lee , J.H, &Hashimoto, H(2004)- Human
Following Mobile Robot In A Distributed Intelligent
Sensor Network. IEEE Transactions On Industrial
Electronics
[7] Burke, M & Brink, W(2010) Estimating Target
Orientation With A Single Camera For use in a Human
Following Robot, South Africa
[8] Gu, J, Ding ,X ,Wang, S, wu : Action Recognition from
Recovered 3-d Human Joints, systems, Man, and
Cybernetics(2010)

Figure 6 : Overall System Design

© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1637