2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON)

Animatronic Hand Model on the Basis of ESP8266

Illya Starodubtsev Igor Gajniyarov Rustam Samedov
IMM UB RAS IMM UB RAS IMM UB RAS
Ural Federal University Yekaterinburg, Russia Yekaterinburg, Russia
Yekaterinburg, Russia i.m.gainiarov@imm.uran.ru rustam.samedov.99@mail.ru
StarodubtsevIS@robolab.tk

Alisa Sibogatova Irina Antipina Yana Zolotareva

Ural Federal University Ural Federal University Ural Federal University
Yekaterinburg, Russia Yekaterinburg, Russia Yekaterinburg, Russia
as2010alisa@gmail.com irinaant_99@mail.ru yanazolotereva@gmail.com

Abstract—3D-printing technology puts the question of For these purposes, there are solutions offered by
augmentation in the fields of rehabilitation, feedback provision KUKA and FANUC companies.
and interaction with real objects into a new perspective. Rapid
Publicly available are open-source models, which can be
and affordable manufacturing of both prototypes and industrial
elaborated for in-house needs. For instance, LIME, InMoov,
designs leads to the emergence of new fields of 3Dprinting
Brunel Hand, Ada Hand. In this paper, we use an InMoov
technology mass use. For example, there already exist hand
[10] hand model for reference.
prostheses targeted at children or animatronic models for
communication. Each of these tasks raises the question of Our stand is suitable for educational purposes as it has a
controlling a physical hand structure. This paper presents an set of well-studied movements. It can be used for training
anthropomorphic hand, which is stand mounted. The focus is on sessions in schools, colleges, universities and other
the software solution for gestures simulation. A special gesture educational institutions for the purposes of studying the
format was developed to tackle this problem. A debugging basics of mechanics, animatronics and algorithms used in
prototype was developed by modifying an InMoov open hand motion capture MEMS. This works in the following way: in
model. The paper presents the original model part as a circuitry motion capture there is gyroscope zero rate error and
and a 3D stand model. The issue of anthropomorphic limb accelerometer flutter, which are offset by different
control is universal. The problem is most pressing in systems algorithms; for example, the Madgwick algorithm. The stand
with fine interaction. Our model covers this problem field. produces reference data, which can be checked both visually
and with software. By comparing the visualization of a
Keywords — animatronics, 3D-printing, augmentation, gesture and the reference data one can fine-tune the
feedback, gestures algorithms of gesture visualization.
I. INTRODUCTION II. PROBLEM STATEMENT
In this paper, augmentation implies physical devices that The existing hand models are represented from models
allow either interacting with the physical world in a large attempting to copy hand biomechanics with maximum
number of ways or acquiring more versatile information accuracy, to industrial prototypes with atypical mechanics
through simulation of movements and gestures. but a larger number of mobility degrees.
Augmentation covers the majority of existing spheres of
human activities. For example, medicine, animation, Animatronics is mainly used in such spheres as medicine,
communication and manufacturing industry. entertainment industry and communications. The following
fields of implementation can be distinguished in a more
By way of specific examples, the following detailed view:
developments in different spheres can be mentioned. In
surgical interventions, these are robotic assistants like • Animation of game and art characters.
Omnicell M5000 Introduction or da Vinci robotic surgeons. • Human limb prosthetics. In this field, cosmetic,
They represent subject matter solutions. In the field of traction and bioelectric developments can be
prosthetics, these are YouBionic and Motorika [6, 7]. The distinguished.
distinctive feature of pediatric prosthetics is the fact that
children grow out of their prosthetic devices fast. Apart from • Rehabilitation assistance.
that, production time and personalization are key parameters.
These are best solved with 3D printing technologies. In the • Use of industrial robots with a versatile, complex and
field often changing activity profile. They can be used in a
human-hostile environment without revamping the
The work was made with the financial support by Decree production line itself.
№ 211 of the Government of the Russian Federation,
contract № 02.03.21.0006. of communication[1], these are In all the cases mentioned above there are a number of
developments for teaching through demonstration [4, 8] or issues, which anthropomorphic robot developers have to
for communication through sign language [5, 12]. In face:
manufacturing industry, these are manipulators for moving, • Managing data deficit. This problem can be solved by
assembling, welding, woodworking and other works. developing a pre-created set of actions, which the end

978-1-7281-4401-6/19/$31.00 ©2019 IEEE 0500

 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON)

device can do. In the case of prosthetics, retrieving The servo drivers and ESP8266 are connected parallel to
information from residual muscles is considered to be the power unit. There are voltage-adjusted controlling flow
the most popular option. However, this only disguises lines between them. In order to imitate human flexorextensor
the problem. muscles adducent strings are used. The material has the
smallest friction factor and weight among its counterparts.
• The discrepancy between the implemented mechanics
of the produced model and the biomechanics of a
hand. The industrial option has deviations towards
more comfortable performance of the preset
operations with a larger or a smaller number of
degrees of freedom. In the case of prosthetics, this
issue is even more pressing, because for end users a
more natural option would be the one accurately
copying the original hand biomechanics.
• Small size. In the case of industrial performance, this
issue does not really burden the developers because
here the device can be stationary in sufficient space.
A factory is designed with a production line in mind.
In the case of prosthetics and communications, the
situation is opposite. There are significant limitations
to the weight and volume of the development.
• Setting device components in motion. A lot depends
on the exact way the limbs are set in motion. For
example, on the type and the number of engines, on
the number of adducent mechanisms.
• The issue of sensing is non-existent in the industrial
sector. However, in the case of prosthetics, the
precision of controlling the prosthetic device depends
on it. The higher quality is feedback from the
prosthetic device to the person, the better the person
can control it [9]. Each of the presented problems
influences the development of control flow [2, 3, 14].
This is what the existing development is focused on.
An InMoov hand model was elaborated as a
prototype to fine-tune anthropomorphic hand gesture
control.
III. TECHNICAL IMPLEMENTATION
Fig. 2. Diagram of connection of servo drivers to NodeMCU
Figure 1 presents the set of gestures that were chosen.
Control is exerted through 5 MG945R servo drives with In comparison with the original hand model, we are
maximum drive torque of 12 kg/cm, a NodeMCU v3 plate offering an addition in the form of rubber finger extensors.
with a Wi-fi ESP8266 module, a power unit with the This is necessary due to the fact that there is a simplification
overcapacity of 300W, and adducent strings. The overall of the extensor mechanics. Previously, the same fishing line
connection diagram is presented in Figure 2. Operating was used for these purposes. Because of this, assembling and
voltage is 5V. pulling of fishing lines became much more complicated. The
overall reliability of the construction was lower compared to
the current solution
After choosing and calculating the element base
modelling stage was carried out. A 3D model of a hand
belongs to the InMoov project, while the model of the stand
itself was designed independently in line with the following
needs: structural stability, easy access to the subsystems and
vivid demonstration of the performance of all the moving
elements.
Modelling was carried out in OnShape. The overall
printing time was about 30 hours. The model of the stand can
be seen in Figure 3.

Fig. 1. Simulated gestures

978-1-7281-4401-6/19/$31.00 ©2019 IEEE 0501

 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON)

Performance tests were conducted to check the servo

driver rotation correspondence with flexion and extension of
each finger.
The results of these tests indicated the necessity of
adjusting the rotation angle of the servo driver within the
task executive program. We were able to set the real rotation
angle for the servo driver and to make the adjustment within
the program.

Fig. 3. A 3D model of the stand with servo drivers

The assembled prototype of an animatronic hand can be

seen in Figure 4. All the elements were printed on an FDM
3D printer. The printer has printing fidelity with the minimal
layer thickness of 100 um, positional precision along the Z
axis of 4 um, and 12 um along axes X/Y.
Printing was carried out in PLA plastic, because this type
of plastic has sufficient flexing strength: about 70−80 MPa at
20% filling.
IV. SOFTWARE IMPLEMENTATION
The Node MCU v3 plate programming was carried out in
the ArduinoIDE development environment. A standard
Servo.h library was used to control the servo drivers. Fig. 5. The pseudocode of the function of controlling the animatronic hand
An algorithm was developed to implement finger finger movements
movements. End positions of servo drivers are sent to the
Here:
input. Each entry is responsible for the movement of a
certain finger. • install_position_time – time of setting the hand
This work algorithm is based on the dependence of the position.
servo driver angle position difference from the fixed time of • rotation_time – overall time of the servo driver
adjustment of the end position of each finger. The software rotation.
performance algorithm in the form of a pseudocode can be
seen in Figure 5. • steps_number – number of steps of the servo driver.
• one_step_angle – one-step rotation angle of the servo
driver.
• new_hand_angl – new hand position.
• hand_angl – pervious hand position.
• servo_rotation – the function of rotating the servo
driver by a preset angle.
In order to imitate smooth movements of fingers it is
necessary to smoothly rotate the servo driver shaft. This kind
of rotation is possible due to quick and frequent change of
the angle [13]. At each iteration the servo driver turns by a
preset angle, which cannot exceed the maximum angle.
The maximum angle is 1° to 5° degrees. The number of
iterations depends on the overall time of adjustment of the
hand end position and on the time spent on one iteration. All
these magnitudes are constant.
V. FUTURE RESEARCH
Upcoming developments include human hand anatomy
research in terms of biomechanics. For instance, the
imitation of finger flexor/extensor muscles will be more
precise and controllable [11], or the imitation of finger
adductor/abductor muscles, or lumbrical and interosseous
muscles.
In order to control an animatronic prosthetic device the
implementation of an interactive interface is planned, which
Fig. 4. Animatronic hand on the stand

978-1-7281-4401-6/19/$31.00 ©2019 IEEE 0502

 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON)

will allow the user to create their own gesture on a virtual [3] E. I. Vorobev, A. K. Skvorchevskij, “Problems of creation of control
hand model. algorithms for anthropomorphic bioprostheses of hands and feet”,
Medicine and high technology, vol. 1, 2012, pp. 7–12 [Medicina i
vysokie tehnologii, p. 7–12, 2012].
VI. FUTURE RESEARCH
[4] L. Zeldin, “Development of movement in various forms of cerebral
At this point in the development of the prosthesis, palsy”, Terevinf, 2014.
animatronic hand control has been implemented in such a [5] A. S. Lobanov, “Overview of gesture recognition techniques”, South
way as to check the feasibility of various positions of the Ural scientific readings, vol. 1, 2015, pp. 48–53 [Juzhno-Uralskie
fingers. In order to demonstrate the prosthesis simple nauchnye chtenija, p. . 48–53, 2015].
gestures were chosen which do not require the work of [6] I. V. Pervakov, K. R. Semensova, “Children's prosthetics: technology
shoulder and forearm muscles, as well as the first dorsal and cultural aspects”, Cultural journal, vol. 4, 2018
[Kulturologicheskij zhurnal, 2018].
interosseous muscle. Besides, due to the peculiarities of the
[7] I. I. Cheh, “Functional prosthetic hands for everyday tasks”,
printed hand model the fact that fingers cannot crook by over unpublished.
90° in the «joints» was taken into consideration. [8] C. Alessandro, “Motor control and learning theories”, Emerging
The little finger and the ring finger can bend back by 45° Therapies in Neurorehabilitation II. – Springer, Cham, 2016, pp. .
225-250.
due to the imitation of metacarpal bones. The proximal
[9] J. T. Belter, “Mechanical design and performance specifications of
phalanx of the thumb allows it to crook by 65° and to bend anthropomorphic prosthetic hands: A review”, Journal of
back by 10° from the hand plane. The input of data Rehabilitation Research & Development, vol. 5, 2013, pp. 50.
concerning the position of each finger is carried out through [10] L. Gael, “Open source 3D printed life-size robot”, unpublished.
a keyboard, i.e. rotation angles for the servo drivers are set. [11] A. Gupta, “Wireless animatronic arm”, in press.
In each final position it is clear by what angle a servo driver [12] D. Homburg, “RoboTalk-Prototyping a Humanoid Robot as Speech-
is rotated at the moment, due to the servo driver feedback. to-Sign Language Translator”, Proceedings of the 52nd Hawaii
International Conference on System Sciences, 2019.
REFERENCES [13] S. F. Keadr, “Development of Zigbee Based Tele Operated
[1] M. Arutjunjan, “Anthropomorphic robotic brush model for Multipurpose Robotic Arm with Hand Gesture Recognition”,
communicative interaction”, Scientific works of the Central Research Technology, vol. 8, 2017, pp. 1275–1286.
Institute of Russian Sign Language, 2018, pp. 133–145 [Nauchnye [14] X. Wang, J. Zhao, D. Yang, “Biomechatronic Approach to a Multi-
trudy Central’nogo nauchno-issledovatel’skogo instituta russkogo fingered Hand Prosthesis”, Proc. of 2010 3rd IEEE RAS & EMBS
zhestovogo jazyka, p. 133–145, 2018]. Inter. Conf. on Biomedical Robotics and Biomechatronics, Tokyo,
[2] E. I. Vorobev, V. O. Dorofeev, A.V. Miheev, “Anthropomorphic Japan. Sep. 26-29, pp. 209—214.
biorobots and bioprostheses”, The Engineering Gazette, vol. 6, 2015, .
pp. 9 [Inzhenernyj vestnik, p. 9, 2015].

978-1-7281-4401-6/19/$31.00 ©2019 IEEE 0503