2020 International Conference on Computational Performance Evaluation (ComPE),

North-Eastern Hill University, Shillong, Meghalaya, India. July 2–4, 2020

Design, Implementation, and Controlling of a

Humanoid Robot
Kiran Somisetti Khushboo Tripathi Jitendra Kumar Verma
Computer Science & Engg. Department Computer Science & Engg. Department Computer Science & Engg. Department
Amity School of Engg. & Technology Amity School of Engg. & Technology Amity School of Engg. & Technology
Amity University Haryana Amity University Haryana Amity University Haryana
Gurugram (Manesar), India Gurugram (Manesar), India Gurugram (Manesar), India
Email: [email protected] Email: [email protected] Email: [email protected]

Abstract—In rapid change with technology a huge development Then he discovered the term robotics in the story Run-Around
of robots is competing equally with humans. The technology in 1942. Isaac Asimovs contribution to the robots played a
development has brought many changes in the field of manufac- vital role in the field of robotics leading to the three laws
turing and research. Initially, the robots were used by humans
to make their work easier, but today the robots are developed to of robotics. Further there were many scientists namely Alan
replicate humans and substitute the labor in the fields where even Turing, George Devo, John McCarthy and Marvin Minksy who
human presence is difficult. The development doesn’t only take contributed intelligence into a robot that grew up under a name
place in the manufacturing of robots but simultaneously it focuses as artificial intelligence today. Now, in the information age,
on control and optimization of the tasks that are done by the there is a huge development of robots. In 2002, the ASIMO
robots. The aspects of robots are unlimited in the field of medical,
defense, education. The type of application defines the robots robot was the most advanced version which could run-walk
construction and the requirement of systems. This paper covers and communicate. Then in 2016, Sophia robot which is a
background of the robotics such as variety of robots, kinematics humanoid robot was developed by Hanson robotics. These
of robots, localization and mapping scenario, the trajectory of continuous developments in the field of robots which artificial
the robot’s path. Also, the control systems used in robots like intelligence embraced in it would take them to greater heights
actuators, sensors and processing of the actuators, robots learning
has been presented. The experiment with humanoid robots for leading to the imagination age of civilization.
civilians has designed through programming and controlled by Some of the most popular existing robots are the biman-
microcontrollers for analyzing the results of degrees of freedom ual underwater humanoid robot that was developed by the
and positions of robots. University of Stanford which was majorly used to reach
Index Terms—Trajectory, Design, Humanoid robots, program- depths that human couldn’t make it reach. These robots were
ming.
used to explore coral reefs of the ocean [1]. It was fully
waterproofing electronic body and its design was featured with
I. I NTRODUCTION
stereoscopic vision. The other type of robot was ATLAS which
A machine which makes the work easier for the human can was designed in the year 2013 by Boston Dynamics with
be called a robot. These machines are programmed and are funding from DARPA that could move through rough terrains
capable of fulfilling the needs of the programmer. Not neces- and obstacles, this robot was able to leap from one platform
sarily a robot must look like a human, rather a simple machine to another and perform backflips too.
which makes our work simpler can be called as a robot. Robots Nao, a very popular robot which is just 23 inches of height
can be either be autonomous or semi-autonomous based on played a major role in the field of education that attracted many
the application so that the user is intended to develop to fulfill people towards it [2]. It was able to perform many operations
his needs. The study of technology which is used to design, like dancing, play a soccer match etc. later RIKEN and
manufacture, assemble and control a robot using sensors and Sumitomo Riko company limited came up with a robot name
actuators is known as Robotics. On a broader note, Robotics Robear which was considered as the solution to the shortage
is the study of robots. From the ancient age of civilization of caregivers. These robots were integrated with important
until the information age (i.e. till the 21st century), there has sensors like rubber capacitance type tactile sensors which
been a huge development in the field of automation. The first could generate gentle movement such as carrying or lifting
robot refers back to the dates back to ancient civilization. The patients without causing any discomfort. Humanoid robots are
father of robots is known to be Al-Jazari who developed a helpful in managing public healths without exposure human
machine in the year 1206 which was an automated machine. being to the deadly infections from bacteria and viruses [3] [4].
Later, in 1921 the word robot was used by an author Karel Recently, in a study it has been found that a humanoid robot
Capek from the Czech word robots means to be a salve. Later was developed for patient’s need protecting from COVID-19
on, the developments were taking place at a smaller scale until due to possible contamination of robots from virus which may
the 1940s when Isaac Asimov introduced stories about robots. be infect as patients and spread virus among other patients [5]

978-1-7281-6644-5/20/$31.00 
c 2020 IEEE
831

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.
 push the incomes of advertising section of the humanoid robot
showcase.
II. H UMANOID ROBOT
A. Background
To understand the concepts of robotics in detail one has
to study the robot initially from the design of the robot until
the movements and computer vision of the system. The initial
design of any robot will be based on the application point
of view. But in general, any design will require few main
Fig. 1: Depiction of Global Market Value of Humanoid Robot things like the degrees of freedom (DOF), joints, torque, and
trajectory planning and control vision that a robot needs to
perform the task required by the user.
DOF can be explained by taking a rigid body in the
[6] [7] [8] [9]. The other major invention in the field of robotics space can be defined by the independent motions. These
was pepper robot which could read human emotions and was independent motions can be described as Translational motion
to serve as a human companion who could communicate and Rotational motion along different directions in x, y and z-
naturally. This type of robot can interpret the tone of the voice axis. The number of independent motions a body can possess
and could respond immediately. This had multi-directional is known as the degree of freedom. The number of restraints
wheels which could make its mobility very easy. The latest can never be zero or six. If the degree of freedom of the body
innovative development was Sophia which was modeled by is zero then it means that there is a joint disconnected and if
Hanson robotics. The author was developed in the year 2015 the degree of freedom is six then it means that the body is
and was granted citizenship in Saudi Arabia. But leaving all solid. To study the structure of the robot, one should first try
these aside, the major challenges of these robots had been to relate them to the living body. The mechanical design of the
about the fabrication methods and the materials to be used robot has many similarities to the skeleton of the human body.
up to build a robot to make it efficient. Not only these but Therefore the robot anatomy would be to study the skeleton
the control systems which had to perform like human and that is in the form of manipulator structure.
have good perception and computation was one of the major The structure of the manipulator comprises of rigid frames
drawbacks. Sometimes finding out the better power sources for that are interlinked to each other employing joints which form
long-lasting and are affordably created a major problem. Poor an arm that safeguards the mobility as well as reachability.
understandability of mobility creates another issue which was These manipulators are fixed on to base which is fixed to
to be focused on a robot to make the design complete robust the floor. The wrist is one of the components that grants the
[10] [11]. To overcome these challenges, we tried to design orientation which performs the task required by the user or
and demonstrate a humanoid robot which is cost-effective and the programmer. The structure of the robot consists of the
efficient with a simple design and has an easy implementation. links that are the rigid bodies of the structure. A body or
Topographically, the worldwide humanoid robot market can group of bodies which are connected and are constrained by
be divided into North America, Europe, Middle East and their relative motion are known as links [4]. These links can
Africa, Asia Pacific, and South America. Asia Pacific is relied be classified into three types based on their ends as binary,
upon to contribute a generous income offer to the humanoid ternary and quaternary links. The links have no relative motion
robot showcase during the figure time frame attributable to between the joints within a link.
developing geriatric populace in nations, for example, China, As like human, there are joints present in a robot skeleton
India, and Japan. also. These joints situated between the links make the robot
The worldwide humanoid robot market can be sectioned to move with ease. The movement of the input and output
based on part, application, and district. In light of part, the links are managed by the joint. Combination of these links and
humanoid robot market can be isolated into equipment and joints make up the robot. For the mobility of robot, motors
programming as in given Fig. 1. The equipment section are used that are connected at the joints and these motors
of the humanoid robot market can be additionally grouped can be programmed based on any microcontrollers [12]. For
into sensor, actuator, controller, power source, and others. each of the microcontroller, it should be programmed such
Ascend in reception of computerized reasoning calculations that there is a code running behind the working of the system.
in humanoid robots to support robots’ exhibition is required This programming can be done in many ways [13]. There are
to an essential explanation behind the main piece of the pie of various key areas where these humanoids can be applied [14]
the product portion. As far as application, the humanoid robot [15].
market can be arranged into individual help, instruction, search Applications. Human-robot interaction is now applied in
and salvage, amusement, inquire about and space investigation, almost all robot tasks, including manufacturing, space, avia-
advertising, and others. Increment in execution of robots in tion, undersea, surgery, rehabilitation, agriculture, education,
retail outlets to improve client experience is relied upon to package fetch and delivery, policing and military operations.

832

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.
 Fig. 2: Circuit Model of the Robot System

B. Mathematical Expression
1) Trajectories of Point to Point Motion: As we know that v0 = a1 + 2a2 t0 + 3a3 t20 (10)
the trajectory that connects to initial and final point while
satisfying other specified constraints at the end points. We
suppose that at time t0 the joint satisfies as shown in (1)–(2). q0 = a0 + a1 tf + a2 t2f + a3 t3f (11)

q(t0 ) = q0 (1)
vf = a1 + 2a2 tf + 3a3 t2f (12)
q̇(t0 ) = v0 (2)
The equations (9)–(12) can be formed into a system of
and for the values at tf , equations as follows.

q(tf ) = qf (3) ⎡ ⎤⎡ ⎤ ⎡⎤
1 t0 t20 t30 a0 q0
⎢ 0 1 2t0 2 ⎥⎢
3t0 ⎥ ⎢ a1 ⎥ ⎢ v0 ⎥
⎢ ⎥=⎢ ⎥ (13)
q̇(tf ) = vf (4) ⎣ 1 tf t2f t3f ⎦ ⎣ a2 ⎦ ⎣ qf ⎦
0 1 2tf 3t2f a3 vf
If we wish to specify initial and final accelerations, then we
need additional two more equations which are as follows. Based on the constraints and the values of the variables and
substituting the values in the above system of equations, one
q̈(t0 ) = α0 (5) can find the values of the coefficients in terms of q0 , v0 , qf ,
and vf .
q̈(tf ) = αf (6)
2) Cubic Polynomial Trajectories: The one way to obtain
a smooth curve is by polynomial function so let us consider
the cubic polynomial,

q(t) = a0 + a1 t + a2 t2 + a3 t3 (7)
Then the velocity can be obtained by differentiating with
respect to time is as follows.

q̇(t) = a1 + 2a2 t + 3a3 t2 (8)

From above two equations, the four equations can be derived
due to the four constraints that are initial position, initial
velocity, final position and final velocity.
Fig. 3: Model of the Humanoid Robot in SOLIDWORKS
q 0 = a 0 + a 1 t0 + a2 t20 + a3 t30 (9)

833

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.
 III. S IMULATION M ODEL AND E XPERIMENTAL S ETUP
To design a robot, the design needs either microcontrollers
or microprocessors (for example, Arduino Uno and Raspberry
Pi), actuators, sensors in terms of hardware and any program-
ming language to implement the code according to the needs
of the user as the software part. Based on the number of
DOFs that many numbers of motors are to be attached to the
controller. Display and camera are inserted in the model to
make the robot more user friendly. Based on the mathematical
formulae, a virtual environment is developed to analyse the
robotic model that was designed by the user in robotics lab.
Initially, the design was developed in SOLIDWORKS with
all the specifications and lengths according to human body
proportions as shown in the Fig. 2. Once the design is
Fig. 4: A Designed Humanoid Robot in Lab
done it is been simulated in V-Rep software where the real-
world environments are defined according to the needs of the
user. Sensors, mechanisms, robots and whole systems can be    
modelled and simulated in many different ways in V-Rep as
shown in Fig. 3. 

The major applications are hardware control, simulation is

very easy, it ensures safety monitoring system. The major




elements of the simulator are scene objects, calculation models     


    
and control mechanisms. The scene objects consist of different

joints, cameras, references planes and many more that can be
used to define the model. Whether the motion should take


place in inverse or forward kinematics, and the minimum
distance to make sure that one can avoid obstacle, collision can


be calculated and programmed accordingly [15]. Path planning
   


can also be defined with the proper control mechanisms. The
experiment parameters considered by us are mentioned in   
Table I.
Fig. 5: Joint Centre Trajectories
IV. I MPLEMENTATION AND R ESULTS
The choice of several motors depends on the DOF that we
want to provide for the robot. As explained above, initially on the real robot takes some more trial and error steps to adapt
the robot must be designed according to needs of the user to the real-world environmental situations to achieve proper
with the minimum number of DOF. Usage of the material accuracy. The results for trajectory planning of the robot have
depends on the size that is needed of designing, material been plotted on the graph as shown in Fig. 4 and Fig. 5.
selection can range from plastic to carbon fibre based on Fig. 4 shows the result of change of hand joint position
our fabrication availability and assemble them. Robot was of the robot with respect to the time. The left hip joint has
fabricated using 3D printing technology with PLA plastic been recorded from 0 seconds to 10 seconds and in mean time
[16]. Once the fabrication is completed assemble the parts the movement takes place from .87 cms to reaches the same
by attaching motors and other electronics in respective places. position with an increase of .02 cms in between at 4 seconds.
After assembly starts giving circuit connection between servo This trajectory shows the movement of left hand from lower
motors, micro-controller, microprocessor and other electronics. position to higher position and then comes backs to the lower
In this robot, all the actions generated by the servo motors are position. In similar way, the right hand moves from higher
controlled by the Arduino. position to lower position and then comes back to higher
The range of motion of each joint can be adjusted using position. When these hands move simultaneously, one hand
the servo movement range in the Arduino program. Raspberry goes up while other hand goes down and vice versa. In this
pi will be sending the commands for Arduino to control the way the centre of gravity has been maintained all over the
motion of robot based on the sensory data coming from the body so the structure fails from falling down.
camera, microphone. Raspberry pi further performs computer Fig. 5 shows the result of change of position of each joint for
vision task with the help of a camera and also helps the robot taking single step. This range of signals is given to the robot
to interact with humans through speech with a microphone and in a cyclic form so that the robot can take multiple steps. The
speaker. Proper stimulation of the robot helps us the tune our above figure depicts the right and left leg joint specifically
joint parameters better and faster. Testing the simulation results knees and ankles. The graph depicts the movement from 0

834

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.
 TABLE I: List of Specification of Parameters and their Functions
Parameters Specification Function
1. Power supply 12v and 5 v 12v used for motors and 5v used for con-
trollers
2. Number of motors Servo motors (15 Nos) Used for the movement of the parts
3. Micro controllers and processors used Arduino UNO and Raspberry Pi Arduino is used for motor controlling and
raspberry pi used camera and speech recog-
nition
4. Servo motor driver Adafruit servo motor driver For controlling the motors
5. LCD display TFT 3.5 inches To display the gestures
6. Speaker and microphone Majorly used for interaction between the
system and the environment
7. Camera Raspberry Pi camera Used for vision

    

     
    











    





Fig. 6: Time Vs. Feet Cycles

Fig. 7: Degree of Freedom Related to Sample Size (n-1)

seconds to 15 seconds. This trajectory shows the movement
of left leg and right leg simultaneously from lower position to
higher position and back to same position [10] [17]. In this
way the centre of gravity has been maintained all over the
body so the structure fails from falling down.
In general, Robot arms are depicted by their degrees of
opportunity i.e. freedom [18] [19]. This number commonly
alludes to the quantity of single-hub rotational joints in the
arm, where higher number demonstrates an expanded adapt-
ability in situating an instrument. Degrees of freedom, in a
mechanics setting, are explicit, characterized modes in which
a mechanical gadget or framework can move. The quantity of
degrees of opportunity is equivalent to the absolute number
of free relocations or parts of movement. The term is broadly Fig. 8: Positions Vs Detection of Motion (Degree)
used to characterize the movement capacities of robots. There
are six total degrees of freedom. Three correspond to rotational
movement around the x, y, and z axes, commonly termed
pitch, yaw, and roll. The other three correspond to translational value towards the mean.
movement along those axes, which can be as moving forward The degree of freedom in the robotics is the main param-
or backward, moving left or right, and moving up or down. eter to calculate in different scenarios. When the humanoid
In order to get the DOF for the estimate, subtract 1 from the robots were designed with experimental specifications after
number of items. Imagine where 4 people, giving 3 degrees implementation in iterative way at different positions detection
of freedom (4 1 = 3), or you could use one hundred people of movement has been captured and depicted as in Fig. 8.
with DOF = 99, where “n” is the number of items in the In figure different symbols are used for iterations 1, 2, 3, 4
data set. Fig. 7 depicts the DOF related to sample size. If and 5. The linear line represents the best fit of motion with
DOF increases this means sample size is increasing. The mean the given points of trajectory during robot motion. During
value lies around 40 in distribution curve. The critical values experiments errors at some time has been recorded and again
decrease while DOF increases due to nature of distribution with correction samples have been taken for consideration of
of points w.r.t. sample size. T distribution pushes the critical path motion by the robot.

835

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.
 V. C ONCLUSION [12] M. W. Spong, S. Hutchinson, and M. Vidyasagar, “Forward kinematics:
the denavit-hartenberg convention,” Robot Dynamics and Control, 2nd
The paper concludes the implementation and demonstration ed.; John Wiley & Sons: New York, NY, USA, pp. 57–82, 2004.
of the humanoid robot with more than 6 degree of freedom [13] R. J. Schilling, Fundamentals of robotics: analysis and control. Simon
which was designed in the SOLIDWORKS software. The & Schuster Trade, 1996.
[14] S. Se, D. Lowe, and J. Little, “Vision-based mobile robot localization
experiment was implemented and tested for 15 DOF. The and mapping using scale-invariant features,” in Proceedings 2001 ICRA.
system was developed according to human body proportions IEEE International Conference on Robotics and Automation (Cat. No.
which produces human expression through the LCD display 01CH37164), vol. 2. IEEE, 2001, pp. 2051–2058.
[15] T. Nguyen, L. Tao, and H. Hasegawa, “The effect of foot structure on
and also recognizes the human or objects with the help of locomotion of a small biped robot,” in MATEC Web of Conferences,
camera which is in its visual range. The implementation of vol. 95. EDP Sciences, 2017, p. 08010.
computer vision was done in the python environment with [16] D. H. Rosenzweig, E. Carelli, T. Steffen, P. Jarzem, and L. Haglund, “3d-
printed abs and pla scaffolds for cartilage and nucleus pulposus tissue
predefined library functions and robot control. The developed regeneration,” International journal of molecular sciences, vol. 16, no. 7,
prototype of hominoid robot is able to perform the set of pp. 15 118–15 135, 2015.
operations very efficiently. In order to increase the robot [17] P. Karkowski and M. Bennewitz, “Real-time footstep planning using
a geometric approach,” in 2016 IEEE International Conference on
performance and control over robot system, higher motor Robotics and Automation (ICRA). IEEE, 2016, pp. 1782–1787.
torque can be used. Thus, the movement of robot will be more [18] L. M. Lix and H. Keselman, “Approximate degrees of freedom tests: A
effective and efficient in the development of robots further in unified perspective on testing for mean equality.” Psychological Bulletin,
vol. 117, no. 3, p. 547, 1995.
real time machines. [19] S. Oßwald, M. Bennewitz, W. Burgard, and C. Stachniss, “Speeding-up
robot exploration by exploiting background information,” IEEE Robotics
VI. F UTURE W ORK and Automation Letters, vol. 1, no. 2, pp. 716–723, 2016.
A variety of approaches exist that tackle the problem of
humanoid locomotion. Among them, the Real-Time Footstep
Planning in 3D Environment is one of the area for research.
Secondly, design of any degrees of freedom, robotic arm and
joints can be taken as re-work and re-think for humanoid
robots. Third, an efficient shortest path finding for mobile
Robots in polygonal maps is also the future work for the
researchers. The mathematics can be more involved for finding
better solutions in terms of trajectory, motion, place, DOF,
error correction schemes.
R EFERENCES
[1] J. Yuh and M. West, “Underwater robotics,” Advanced Robotics, vol. 15,
no. 5, pp. 609–639, 2001.
[2] M. Ben-Ari and F. Mondada, “Robots and their applications,” in Ele-
ments of Robotics. Springer, 2018, pp. 1–20.
[3] G.-Z. Yang, B. J. Nelson, R. R. Murphy, H. Choset, H. Christensen,
S. H. Collins, P. Dario, K. Goldberg, K. Ikuta, N. Jacobstein et al.,
“Combating covid-19the role of robotics in managing public health and
infectious diseases,” 2020.
[4] “Engineers develop robots to treat and test covid-19 patients in a
bid to protect health workers,” https://www.telegraph.co.uk/global-
health/science-and-disease/engineers-develop-robots-treat-test-covid-19-
patients-bid-protect/, (Accessed on 04/20/2020).
[5] A. A. Malik, T. Masood, and R. Kousar, “Reconfiguring and ramping-up
ventilator production in the face of covid-19: Can robots help?” arXiv
preprint arXiv:2004.07360, 2020.
[6] A. Mottrie, “Erus (eau robotic urology section) guidelines during covid-
19 emergency,” 2020.
[7] R. H. Verheijen, “Robot assisted surgery during the covid-19 pandemic,
especially for gynecological cancer: a statement of the society of
european robotic gynaecological surgery (sergs).”
[8] K. D. Stensland, T. M. Morgan, A. Moinzadeh, C. T. Lee, A. Briganti,
J. W. Catto, and D. Canes, “Considerations in the triage of urologic
surgeries during the covid-19 pandemic,” European Urology, 2020.
[9] S. Du, J. Wang, H. Zhang, W. Cui, Z. Kang, T. Yang, B. Lou, Y. Chi,
H. Long, M. Ma et al., “Predicting covid-19 using hybrid ai model,”
2020.
[10] A. Gasparetto, P. Boscariol, A. Lanzutti, and R. Vidoni, “Trajectory
planning in robotics,” Mathematics in Computer Science, vol. 6, no. 3,
pp. 269–279, 2012.
[11] S. Kuindersma, R. Deits, M. Fallon, A. Valenzuela, H. Dai, F. Permenter,
T. Koolen, P. Marion, and R. Tedrake, “Optimization-based locomotion
planning, estimation, and control design for the atlas humanoid robot,”
Autonomous robots, vol. 40, no. 3, pp. 429–455, 2016.

836

Authorized licensed use limited to: Carleton University. Downloaded on September 19,2020 at 06:42:06 UTC from IEEE Xplore. Restrictions apply.