Visvesvaraya Technological University

“JnanaSangama’’Belagavi-590018

Seminar Report on

“Flagellate Underwater Robotics at Macroscale”

Submitted by

Ms. K N DEEPSHI (1VE17EC045)

Under the guidance of

Prof. ABC
Assistant Professor
Dept. of ECE, SVCE

Department of Electronics & Communication Engineering Sri Venkateshwara

College of Engineering
Bengaluru-562157

2021-2022

Sri Venkateshwara College of Engineering

Department of Electronics & Communication Engineering
 CERTIFICATE

Certified that the seminar work entitled “Flagellate Underwater Robotics at Macroscale” was
carried out by Ms. K N DEEPSHI (1VE17EC045) who is a bonafide student of VIII Semester
Electronics & Communication Engineering, Sri Venkateshwara College of Engineering. This is
in partial fulfilment for the award of Bachelor of Engineering in of the Visvesvaraya
Technological University, Belagavi, during the year 2021-2022. It is certified that all
corrections/suggestions indicated for seminar internal assessment have been incorporated in the
report. The seminar report has been approved as it satisfies the academic requirements in the
respect of seminar work prescribed for the said degree.

Prof. Guide Name Dr. Jijesh J J Dr. Nageswara Guptha M

Assistant Professor Associate Professor & Principal
Head
Dept. of E&CE, SVCE Dept. of E&CE, SVCE SVCE
 SRI VENKATESHWARA COLLEGE OF ENGINEERING

DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING

DECLARATION

I hereby declare that the seminar work entitled “Flagellate Underwater Robotics at Macroscale”
submitted in the partial fulfillment of the requirements for the award of the degree of the
BACHELOR OF ENGINEERING, in ELECTRONICS AND COMMUNICATION
ENGINEERING of the VISHWESHWARAYA TECHNOLOGICAL UNIVERSITIES,
BELGAVI is an authentic record of my own work carried out during 2021-2022.The report
embodied in this seminar report has not been submitted to any other university or institute for the
award of any degree or diploma.

PLACE: BENGALURU K N DEEPSHI

DATE: 1VE17EC045

ACKNOWLEDGEMENT

My most sincere and grateful thanks to SRI VENKATESHWARA COLLEGE OF

ENGINEERING, the temple of learning, for giving me an opportunity to pursue the B.E in
Electronics and Communication Engineering and thus helping me to shape my career. First and
foremost, I would like to express my gratitude to Dr. NAGESWARA GUPTHA M, Principal of Sri
Venkateshwara College of Engineering, Bengaluru, for his support in bringing this seminar to
completion.

I would like to extend my sincere thanks to Dr. JIJESH J J, Head of Department of Electronics and
Communication Engineering, Bengaluru, for his suggestions which helped me to complete the
technical seminar.

I would like to express sincere thanks to guide Prof. Guide Name, Assistant/Associate/Professor,
Department of Electronics and Communication Engineering, Sri Venkateshwara College of
Engineering, Bangalore, for guidance and support in bringing this technical seminar to completion.

Yours Sincerely,
K N DEEPSHI
(1VE17EC045)

TABLE OF CONTENTS

Chapter. No. Contents Page No.

1. Introduction 1
2. Literature Review 2
3. Methodology 3
4. 6
5. Results and Discussion 7
6. Conclusion 8
7. References 9
 Title of seminar 2021-2022

Chapter 1
INTRODUCTION

ROBOTICS is pushing forward the boundary of exploration, aiming toward remote planets in
space and deep waters within our oceans. Either for commercial or scientific purposes, humans
depend on robotic intervention to operate in such extreme environments, where pressure,
temperature, and lack of illumination prevent scuba diving and human presence. The role of
robotics is predominant with respect to human intervention also in shallow water: longer
operation times, resilience to atmospheric conditions, and reduced human risk are just a few
among the many advantages that motivate the employment of robots over humans. However,
remotely operated and unmanned underwater vehicles have been introduced to replace human
divers in underwater operations with incomplete success. While they have been successfully
applied to safely patrol open stretches of sea, they still present severe limitations in manipulation,
close structure monitoring, and locomotion within cluttered environments. In particular, they
suffer from no negligible limitations when navigating close to the seabed or near submerged
structures, where unintended impacts must be prevented consistently. For all these reasons, some
researchers suggested a radically different approach toward the soft robotics paradigm. Soft
robotics appears, both in land and especially in water, as a promising alternative to allow safer,
simpler, and cost-effective interactions with the environment. The increasing interest in the use
of compliant and soft material stirred the development of several underwater soft robots, which
exploited the compliance in their body to increase their own performance, or to grant new
capabilities to underwater vehicles.
Within the context of aquatic organisms, major sources of inspiration have been drawn from fish
and cephalopods [1]– [7]. Looking at swimming microorganisms, a crucial source of inspiration
for soft propellers comes from prokaryotic flagellum [8] [9], which is considered as the only
known example of a biological “wheel” [10]. Its simple rotatory propulsion mechanism makes it
similar to traditional propellers and more practical with respect to other existing soft underwater
robots based on oscillatory motion. A great number of flagellum-inspired prototypes exists,
characterized by a stiff structure and ranging in size in the micrometer scale [11], operating in
the low-Reynolds number regime. In [12], artificial swimmers with flexible tail shown nontrivial

Dept. of E&CE, SVCE, Bengaluru Page 6

 Title of seminar 2021-2022

dependence of swimming speed and direction on actuation frequency under a rotating magnetic
field, but these were also ranging at the micrometer scale. A macroscopic experimental
investigation of the elastic trail propulsion mechanism is presented in [13], where a high
viscosity fluid is considered to restrict the analysis at the Low-Reynold number regime. A soft
macroscale propeller inspired by prokaryotic flagella has been recently proposed in [14].
In this previous work by the authors, the design of the prototype has been presented and a first
set of trials has been conducted in a controlled environment to assess the propulsive thrust. Here,
we extend and improve the results to propose a first class of macroscale soft underwater vehicles
inspired by flagellated bacteria. In particular, we provide a theoretical model to investigate the
propulsive capabilities of the macroscale flagellar modules, which are hybrid rigid-soft elements
made up of three components. In this article, the final structure resembles the one of the
prokaryotic flagellum, but a great number of solutions can be generated by the combination of
different designs of the three components, paving the way to a new family of soft underwater
robots. Finally, we present an underwater vehicle, which is propelled by combination of four
flagellar modules. This article is organized as follows. Section II presents the bioinspired design
of the flagellar propulsive modules and their fabrication. In Section III, the theoretical model is
presented, with particular attention to the description of the action of the fluid on the modules,
which generates the propulsive thrust. Section IV presents the theoretical and experimental
analysis that have been carried out to complete and validate the theoretical model. Section V
provides two experimental set-ups that are employed to perform a first estimation of the
propulsive capabilities of the single flagellar module. Section VI presents the design, fabrication,
and experimental tests of an underwater vehicle prototype, which, similarly to aerial quadrotors,
combines the action of four propellers to move in different directions. Finally, Section VII draws
the conclusive discussion on the obtained results and on the future developments of the proposed
research.

1.1 Motivation

By centering the above issues, comes the need to create and plan an ease water quality observing
framework that can screen water quality progressively utilizing IoT environment.

1.2 Problem Formulation

Dept. of E&CE, SVCE, Bengaluru Page 7

 Title of seminar 2021-2022

By centering the above issues, comes the need to create and plan an ease water quality observing
framework that can screen water quality progressively utilizing IoT environment.

1.3 Objectives

 To understand the …………..

 To …………………..
 To of the ………….
 To …………….

1.4 Advantages and Disadvantages

1.5 Applications

1.6 Chapter Summary

Chapter 2

LITERATURE REVIEW

In this paper [3] the author had described that the water checking frameworks,

Dept. of E&CE, SVCE, Bengaluru Page 8

 Title of seminar 2021-2022

2.1 Abc

2.2 Chapter Summary

Chapter 3

METHODOLOGY

BIO-INSPIRED DESIGN AND FABRICATION By taking inspiration from the prokaryotic

monotrichous bacteria [15], we designed a novel kind of soft propeller embedding the principal
components of the bacteria’s flagellum (see Fig. 1) and ranging in the macroscale size. The
natural flagellum is made of two main components: the hook, which is the short proximal part of
the flagellum, and the filament, which is the long distal part. An actuation structure, the flagellar
motor, connects the flagellum with the cell’s body and rotates the hook, which eventually
transmits the torque to the filament. The interaction between the elastic component of the
flagellum and the surrounding fluid creates the onset of helical waves [17], which finally propel
the bacteria forward, or backward, depending on the travel direction of the wave [15]. Soft

Dept. of E&CE, SVCE, Bengaluru Page 9

 Title of seminar 2021-2022

mobile robots, thanks to their deformable bodies or appendages, could exploit simple actuation
strategies to achieve complex and rich behaviors [2]. With this respect, the prokaryotik flagella
represents a perfect biological model to exploit the different mechanical properties of the hook
[18] and the filament [19] under the effect of a simple rotational actuation granted by the
flagellar motor [20]. Moreover, the hook shows high resistance Fig. 1. (a) B. parahaemolytica
and (b) the basal discs of the flagella of P. fischeri, [16]. (c) Schematics of the flagellum
structure with the flagellar motor. Fig. 2. Sketch of the flagellum design. to torsional strain,
ensuring the transmission of the torque from the flagellar motor to the filament. These two
features allow a high number of possible kinematics, which would not be possible with a
flagellum made of an homogeneous structural material [21]. Since the final shape of the
flagellum, and thus the locomotion performances of the bacteria, emerges as an interplay
between the fluid, the flagellum mechanical properties, and the actuation frequency of the motor,
it could be preferable to investigate the structural response of such passive elements without a
priori imposing a certain helical shape (i.e., a rigid appendage) [11], [15].

A. Flagellar Module Design and Fabrication:

The propeller’s final design has been first presented in [14]. Following the
abovementioned specifications, the soft module is composed of a cylindrical hook,
having a length Lh = 50 mm and a precurvature αh = 45◦, and a conical filament 300 mm
long, with a diameter varying from 2 to 25 mm (see Fig. 2).

Dept. of E&CE, SVCE, Bengaluru Page 10

 Title of seminar 2021-2022

The flagellum is made out of silicone, while the torsional stiffness of the hook has been achieved
with an ABS backbone, specifically designed and cast inside the silicone. After being vacuum
degassed, the silicone is injected in a Teflon mold that has been fabricated via CNC machining.
Using a Teflon made mold, there is no need to use any release agent, significantly easing the
casting process. The 3D-printed backbone of the hook, including the attachment to the motor’s
shaft, is inserted in the mold before proceeding with the injection (see Fig. 3). Different
parameters and variables affect the net propulsive thrust of the proposed device. These include
the density of the material employed for fabricating the flagellum, the frequency of the motion
and the viscosity of the fluid.

Dept. of E&CE, SVCE, Bengaluru Page 11

 Title of seminar 2021-2022

Dept. of E&CE, SVCE, Bengaluru Page 12

 Title of seminar 2021-2022

Here, we focus mostly on the different performances that can be obtained at varying the
deformed shapes of the propeller, namely the helical waves, which are obtained at varying the
material stiffness and the rotational speed of the motor. For this reason, three different silicones
have been used to fabricate the flagellar modules, all from Smooth-SilTM. Finally, a rigid
flagellum, made out of ABS, has been 3D-printed to compare the behavior that can be attained
using a material that does not allow the onset of the helical waves along the filament. Fig. 4
presents the four flagellar modules, while the main properties of the employed materials are
reported in Table I.

An ad-hoc designed stainless steel shaft is used to connect the hook’s backbone to an MX28-AT
Dynamixel servomotor by ROBOTIS. In particular, with respect to other similar servomotors,
the MX28-AT allows the control of the propeller in wheel mode, maintaining a predefined
constant rotatory speed. The motor is contained in a waterproof container and is controlled by an
Arbotix-M Robocontroller, an Arduino compatible microcontroller specifically made for
Dynamixel actuators. Both the board and the motor are power by a 11.1 V, 2200 mAh Li-Po
battery. Two XBee 1 mW Wire antennas are used to initiate the rotation of the motor and to
continuously track the actuation speed.

Dept. of E&CE, SVCE, Bengaluru Page 13

 Title of seminar 2021-2022

B. DYNAMICAL MODELING:

The elastodynamics of the proposed hybrid soft-rigid design is described through the geometric
and unifying approach developed in [22] and [23], which is a generalization to soft and hybrid
systems of the geometric theory of rigid robots due to Brockett [24]. For further details, the
interested reader is referred to [22]. According to this formulation, the configuration of the ith
soft/rigid body with respect to the spatial frame is defined as a curve gi(·) : Xi → gi(Xi) ∈ SE(3),
given by the following:

where Xi ∈ [0 , Li] is the soft body curvilinear abscissa (Xi = 1 for rigid bodies), Ri(Xi) ∈ SO(3)
is the orientation matrix and ui(Xi) ∈ R3 is the position vector. The configuration space is
parametrized by the traditional generalized variables for the rigid components and by a subspace
of the 6-D local strain space for the soft ones, defined in (2). A schematics of the kinematics of
soft-rigid hybrid systems is reported in Fig. 5. The continuous models of the position, velocity,
and acceleration of a soft body can be derived from the Cosserat rod theory, which gives [25]

Dept. of E&CE, SVCE, Bengaluru Page 14

 Title of seminar 2021-2022

. In the above expressions, the tilde represents the isomorphism between 3-D vectors and skew
symmetric matrices (see Appendix A). To model a constrained rod, a constant strain is assumed
along the section jth, providing

where j>i are body (or section) indexes and Bj ∈ R6×nj is the basis for the subspace of motion

Dept. of E&CE, SVCE, Bengaluru Page 15

 Title of seminar 2021-2022

allowed by the jth joint (or strain) and the body generalized coordinate qj ∈ Rnj (nj being the
number of degrees of freedom of the jth body). ξ∗ j ∈ R6 is the elastic joint (or strain) reference
value, which is equal to zero for lumped inelastic joints. Assuming piece-wise constant strains
[26], equations (2) can be analytically integrated in space using the matrix exponential method,
leading to the kinematics of a hybrid system

where Xj is taken along the jth discrete section. Following the assumption of constant strains,
equation (4)1 provides the jth body position and orientation gj ∈ SE(3) thorough the exponential
mapping of the joint or strain twist. Finally, equation (4)2 provides the jth body velocity twist ηj
∈ R6, where gijgj ∈ SE(3) is the jth body position and orientation with respect to the ith body,
while Adgj and Tξj ∈ R6×6 are, respectively, the adjoint and tangent operator of the exponential
map (see Appendix A). Successive applications of equations (4) to all the bodies of the hybrid
system, yields to the definition of the geometric Jacobian Jj ∈ R6×n for each soft/rigid body (n
being the total number of DoFs), which relates the generalized coordinate vector q ∈ Rn and the
jth body velocity twist ηj [22]

number of bodies. The free dynamic equation is obtained, for a rigid body is

Dept. of E&CE, SVCE, Bengaluru Page 16

 Title of seminar 2021-2022

where Mi ∈ R6×6 represents the screw inertia matrix of the rigid body, FJi and FJj ∈ R6 are the
internal load transferred through joints i and j, respectively, while Fei ∈ R6 is the concentrated
external load. For a soft body, the free dynamics equations with their boundary conditions are

Above, M¯ i ∈ R6×6 is the screw inertia density matrix of the cross section, F¯ ei ∈ R6 is the
distributed external load, F¯ ii ∈ R6 is the internal wrench due to the elasticity of the soft body.
A linear elastic model is employed, since this assumption very well describes the behavior of the
considered rubber-like materials up to 100% elongations:

where Σi = diag(GJxi , EiJyi , EiJzi , EiAi, GiAi, GiAi), E is Young’s modulus, G is the shear
modulus, A is the section area, and Jy, Jz, and Jx are, respectively, the bending and torsion
second moment of area of the beam cross section. Once a Jacobian is found from (5), the
generalized dynamics of the hybrid system can be obtained by projecting the free dynamics of
each body, both rigid (6) and soft (7), by virtue of the D’Alembert’s principle. For further details,
the reader is referenced to Appendix B and to the previous works by the authors [22] and [23].
Thus, we obtain the generalized dynamic equations of the hybrid system in the standard form

Dept. of E&CE, SVCE, Bengaluru Page 17

 Title of seminar 2021-2022

The above expressions are valid for a soft element, while the rigid case can be easily recovered
removing the integrals and replacing the distributed quantities with their lumped counterparts.
Moreover, in general, for a rigid link τ i = i ST i FJi .

C. Hydrodynamical Modeling:
The dynamic interaction between the soft filament and the surrounding fluid represents the
foundation of the propulsive mechanism of the flagellar modules. The filament deforms as a
consequence of the action of the fluid, which, at the same time, is affected by the movement and
rotation of the module. The complete description of such an highly coupled fluid-structure
interaction requires a deep investigation, possibly with the support of CFD software, which goes
beyond the scope of the proposed paper and that will be subject of further investigation in future
works. Here, we propose a simplified model that allows the decoupling of the two problems. In
particular, in the proposed model, the action of the fluid on the ith body is included in the
external force vector as

where we omitted the ith subscript for sake of clarity. The subscripts G, B, A refer to the gravity,
buoyancy, and added mass components of the external force, respectively, while the subscript D
represents the drag and lift components. In the following, all the local reference systems are

Dept. of E&CE, SVCE, Bengaluru Page 18

 Title of seminar 2021-2022

positioned in the geometric centers of the bodies, with the x axis for the flagellar module
corresponding to the axis tangent to the midline, while for the rigid canister the y axis
corresponds to its axis of symmetry.
1) Gravity and Buoyant Force: The external force vector includes the action the buoyant
force and the weight of the body, both acting in the vertical direction in the world
reference system. Following Archimedes’ principle and assuming a uniform density of
the fluid ρw, the buoyant force is directly proportional to the volume of the displaced
fluid and the fluid’s density. This force is applied to the geometric center of the displaced
volume of fluid, which coincides with the entire external volume of the body Vext. For
the soft, continues bodies with uniform mass distribution, the combination of the weight
and the buoyant force is given by

where ρ is the body’s density and G is the gravity vector in the local reference system
(whose rotational components are null). When the weight is not uniformly distributed, for
example, for the canister containing the motor and other electronic components, the
weight and the buoyant force should be calculated separately. In particular, the weight is
obtained through the definition of the body total inertia matrix, given by the sum of the
inertia matrices of all the components, projected in the body’s reference system, namely

where k spans among the weights composing the body, Mk is the inertia matrix of the kth
components in its local reference system and Ad∗ g is the coAdjoint representation of
SE(3) (see Appendix A). Following this procedure, the total inertia mass matrix M
resembles the following structure [27]:

where I ∈ R3×3 represents the rotational inertia of the body, m is its total mass, c ∈ R3 is
the position of the center of gravity with respect to the body’s reference system and I3 is
3 × 3 identity matrix. Once the total inertia matrix is found, the resultant of the weight
and the buoyancy force is

Dept. of E&CE, SVCE, Bengaluru Page 19

 Title of seminar 2021-2022

2) Added Mass: The added mass matrix MA models the volume of fluid accelerated around
the body. It represents the reactive, inertial forces caused by the interaction with the fluid,
which are not negligible in the high Reynolds number domain and when slender bodies
are considered [28]. The action of the added mass is obtained as follows:

The added mass matrices for the different modeled elements are defined as follows:

It should be noted that the off-diagonal terms of the added mass matrix might be non-null
but, as previously mentioned, through a parallelism with rigid blades, these terms are
neglected for sake of simplicity. In (20)1, R(X) ∈ R is the radius of the flagellum in X,
while By = Bz = 0.6 are the added mass coefficients in the y and z direction. For sake of
simplicity, the rigid body is approximated to a symmetric prolate ellipsoid with radius Rb
and total length Lb. Following this approximation, the added mass terms λ, β, and γ in
(20)2 are [29]:

where = 1 − 4( Rb Lb )2 is the eccentricity of the body.

3) Drag and Lift Forces: Regarding the modeling of the fluid-structure interaction in the
external force vector, we employ the simplified Lighthill 3-D large amplitude elongated
body theory [30], where the fluid action is represented by a distributed force along the
flagellum, given by

In (22), ν(X) ∈ R3 is the translational part of the velocity twist η(X), while D(X) ∈ R6×6

Dept. of E&CE, SVCE, Bengaluru Page 20

 Title of seminar 2021-2022

is the screw matrix of the drag and lift coefficients. Through this formulation, the action
of the fluid is decomposed in a resistive (drag) and a conservative (lift) part. While the
drag force varies mostly with the Reynolds number Re of the structure and acts opposite
to the flow direction, the lift force is mostly due to the circulation of the flow around the
body induced by the radial gradient of the velocity (Kutta–Joukowski theorem) and it acts
perpendicular to it. Following the classical aerodynamics theory of propulsive systems,
shear stresses and edge effects are neglected for sake of simplicity. This assumption
would not be valid at low Reynolds numbers, while, in our case, the order of magnitude
of the Reynolds number Re has been estimated to be 104. Finally, the drag and lift matrix
D is modeled as follows:

where CL is the lift coefficient along the x direction (the direction tangent to the cross
section of the flagellum), while CD is the drag coefficient acting on the y and z
directions. The procedure employed to obtain the drag and lift coefficients composing
matrix (23) for the soft and the rigid flagella is detailed in the following Section IV.
Moving to the rigid body, its drag and lift matrix component is obtained as follows:

a) Motion Subspaces for the Flagellum Module and the Rigid Components:
The proposed design is composed by one rigid body (the flagellar motor) and two soft
components (the hook and the filament) having different elastic features. Each of these
three components are therefore described by a specific model, which better represents
their behavior. In particular, the flagellar motor is modeled as an actuated revolute joint,
the hook is represented by an inextensible Cosserat rod with constant curvature and
constrained torsion, while the filament is described by an inextensible Kirchhoff–Love
rod. Accordingly, the motion subspaces are defined by the following basis and reference
twists (assuming the x-axis lying along the rotation axis for the revolute joint and along
the midline tangent for the rods):

Dept. of E&CE, SVCE, Bengaluru Page 21

 Title of seminar 2021-2022

Dept. of E&CE, SVCE, Bengaluru Page 22

 Title of seminar 2021-2022

Chapter 4

RESULTS AND DISCUSSION

4.1 EXPERIMENTAL TEST ON A SINGLE FLAGELLUM

In this Section, two experimental trials are used to provide a first estimation on the swimming
capabilities of the modules. In the first set of trials, the prototype is able to self propel in a marry-
go-round set-up. The results have been fundamental to validate the accuracy of the theoretical
predictions and to assess the propulsive capacities of the modules, confirming the predictions
about the reverse in the thrust direction when going from a soft to a rigid flagellum. In the
second trials, a robot prototype, composed by a single flagellar module, has been designed and
fabricated.
4.11 Self Propulsion Test
In these trials, the flagellar module is constrained by a passive, fixed stand in a merry-
go-round set-up (see Fig. 10). Through a clamp, the waterproof canister containing the
actuating motor is held by a passive arm, which is connected to the fixed stand containing
a 1024 P/R (Quadrature) rotary encoder. Bearings are used to ensure friction-less rotation
of the arm. In this case, the Arbotix-M Robocontroller and the Li–Po battery used to
control and power the motor are placed on top of the passive arm (see Fig. 11). The thrust
provided by the onset of helical waves along the filament is converted in the rotation of
the passive arm around the fixed stand. Using an Arduino Mega connected to a laptop,
the rotatory encoder is employed to read the rotational speed and to relate the propulsive
capability of the flagellar modules to their material stiffness and to the speed provided by
the actuation motors. The trials have been performed at the same four different speeds
used for the statics tests in Section IV-B, both clockwise and counter-clockwise, for 4
min. The swimming speeds obtained during these tests are reported in Fig. 12. Finally,
the results have been compared to those predicted by the theoretical model, in order to
validate its accuracy. Table IV presents the swimming speeds obtained both in the
experimental tests and through the model.

Dept. of E&CE, SVCE, Bengaluru Page 23

 Title of seminar 2021-2022

It can be noted that: 1) at the lowest motor’s speed, ω1 = 0.75 ωmax, the Dragon SkinTM30
silicone, namely the softest one, provides the highest swimming speed; 2) at the highest motor’s
speed, ω4 = ωmax, the SmoothSilTM940 silicone, namely the stiffer one (excluding the rigid
flagellum) provides the highest swimming speed; 3) at all motor’s speeds, the rigid flagellum
provides a negative thrust which, instead of pushing forward the arm, pulls it, resulting in a swim
in the opposite direction with respect to the deformable modules. In particular, the change in the
direction of the thrust when going from a soft to a rigid propeller proves once again the relevance
of elasticity and the complete opposite results that can be obtained using deformable materials.
In particular, the drag and the lift forces are directed, respectively, perpendicular and parallel to
the direction of the flow with respect to the cross section of the propeller. In the rigid case, these
directions are constant along the length of the flagellum and their inclination is set by the
precurvature of the hook, which, in our design, is equal to 45◦. On the other side, as a
consequence of the interaction with the fluid, the soft flagella are highly deformed, resembling
an helical shape whose properties depends on the stiffness of the material and on the rotational
speed of the motor. For this reason, the direction of the fluid flow with respect to the cross
section varies along the length of the flagellum, thus different directions of the drag and lift
forces are observed, finally providing a reversion in the resultant thrust force with respect to the
rigid case. Ideally, a limit value of the stiffness of the module could be found, where the change
in the direction of the thrust is realized. This will be subject of further investigation on a future
work. In conclusion, the obtained results prove that the theoretical model is capable of precisely
predicting the behavior and, in particular, the thrust provided by the modules (both rigid and
soft). Moreover, the swimming capabilities of the flagellar propellers have been proved, paving
the way to the design and fabrication of a new family of robots that rely on these modules to self
propel.
4.12 Single Flagellum Prototype:
In this section, a robot prototype, composed by a single flagellar module, is designed and
fabricated. The same waterproof canister employed in the self propulsion test is now used to
contain the MX28-AT motor. The robot is tethered through particularly flexible and long
wires, which are connected to the LI-Po battery and the Arbotix board placed outside the
pool. The free swimming of the robot has been tested in a 2-m wide and 4 m long pool, with

Dept. of E&CE, SVCE, Bengaluru Page 24

 Title of seminar 2021-2022

a water level depth of 60 cm. Using a single module, particular attention needs to be paid to
constrain the rotations of the canister that are caused by the motor torque. In fact, in Section
IV-B (see Fig. 9), it has been showed that only two reaction forces are not null during the
rotation of the single flagellum: the propulsive thrust fy and the out-of-plane torque my.
Bacterial flagella employ a counter-rotation of the head to compensate the screw rotation of
the tail, resulting in total torque balance. For the sake of simplicity, in the proposed
prototype, a system of boat hull structures has been designed and 3D-printed using ABS in
order to constrain the out-of-plane torque my. In particular, the structure is composed of two
canoe-boats and a system of beams which are finally tightened to the canister. The
submerged part of the hull structure naturally adapts to the motion of the prototype, assuring
the required rotational constraint to the canister for the entire time of the motion. Finally,
some weights have been placed on one end of the canister to assure the full submersion of the
vehicle. The motor has been powered at its maximum speed ω4 = 51 r/min. Some snapshots
of the vehicle during an experiment are reported in Fig. 14. A straight motion is obtained,
with an average swimming speed of 2.5 cm/s. It should be noted that the considered silicone,
in the self-propulsive set-up in Section V-A, presented a swimming speed around 1 cm/s (see
Table IV). The lower swimming speed is mostly due to the presence of the passive arm that
is also pushed by the propeller, while in the single flagellum prototype only the canister is
added.

FLAGELLATE VEHICLE:

A completely untethered vehicle, equipped with four flagella, has been designed, fabricated
and tested (see Fig. 15). An aluminium shell is designed to contain the motors, the battery
and all the required electronic components. The shell is divided in two elongated domes
having a diameter of 180 mm each, which are joined using rubber O-rings and screws to
obtain a total length of 210 mm. The first dome contains an ABS 3D-printed structure
supporting the four motors which, through 4 mm holes sealed with O-rings, are connected to
the flagellar modules outside the shell. The second dome contains the Arbotix-M
robocontroller and the Li–Po battery, which is placed inside a 3D-printed ABS box secured
to dome’s walls to prevent any damage to the electronic components during the motion of the

Dept. of E&CE, SVCE, Bengaluru Page 25

 Title of seminar 2021-2022

vehicle. A magnetic sensor is connected to the Arbotix-M Robocontroller to launch the

motor’s rotation. Finally, some weights have been placed to compensate the buoyancy due to
the air trapped and to obtain a straight at-rest position of the vehicle in the water. In this case,
the four flagellar modules are made out of Mold Star 30 silicone. The position of the four
motors on the shell has been analyzed in order to obtain the maximum thrust preventing, at
the same time, any possible entanglement between the flagella(see Fig. 16). For this reason,
they are positioned resembling the configuration of the propellers on a quadrotor systems and
with an inclination angle of 45◦ with respect to the center line of the dome. Moreover,
opposite flagella present a 180◦ phase shift in their at-rest position and they also rotate with
alternate directions. In this way, the out-of-plane reaction torque, which was compensated by
the hull structures for the single module prototype in Section V-B, now automatically
vanishes as a Fig. 16. Design of the aluminium shell for the four-modules vehicle. The
position of the motors has been specifically determined in order to sum the thrusts provided
by the four flagella and, at the same time, to prevent any entanglement. To obtain a straight
motion, all the four motors are run at the same angular speed, with opposite directions. For a
steering motion, only to motors are run, with opposite direction. In the Figure, c.m.
represents the center of mass of the vehicle. combination of the opposite reactions provided
by each flagella. The theoretical model presented in Section III is employed to first estimate
the performance of the designed vehicle and to investigate the motors speeds to obtain
different swimming modes. The model is expanded to consider the presence of four
propellers and the rigid body is modeled to represent the geometry of the shell and the weight
that it contains. In particular, as mentioned in Section III-A, it is crucial to precisely calculate
the distribution of the mass inside the shell in order to correctly model the distance between
the center of gravity and the center of buoyancy. In the considered case, the total inertia
matrix of the canister in (16) is given by the shell’s inertia, the battery with its case, the
Arbotix-M Robocontroller, the weights that were used to balance the structure in the water
and the four motors with the structure holding them. In this way, from (17) it was possible to
observe that the center of gravity c of the canister is positioned toward the front part of the
vehicle. Considering the presence of the flagella, which are positioned on the back of the
vehicle, the center of the mass of the entire vehicle is thus positioned in its geometrical
center, which is also the center of buoyancy, ensuring the stability of the system and

Dept. of E&CE, SVCE, Bengaluru Page 26

 Title of seminar 2021-2022

preventing unintended rotation.

Straight Motion :
In order to obtain a straight motion, the simulations are run with the four motors rotating
at the maximum speed, with alternate directions. The results of the simulations are
presented in Fig. 17, where some snapshots of the simulations are presented and
compared with the results of the experiments performed with the same specifications. At t
= 0, it is possible to notice the initial configuration, with two modules in a straight
position, while the opposite two present a 180◦ phase shift. The motors then start rotating
with alternate directions and the steering straight motion is attained. After 13 s, the robot
travelled 70 cm, with an average swimming speed of 5.38 cm/s. The power consumption
of the prototypes is mostly related to the consumption of the motors, which is around 22
W. The complete video of both the simulation and the experiment are contained in the
supplementary material. In this regard, the proposed theoretical model represents a useful
tool to efficiently investigate the behavior of different designs and geometries of both the
propellers and the vehicle.
Steering Motion :
Following the approach for the flight control of quadrotor systems, in order to obtain a
steering motion, two motors are run at the maximum speed, with alternate directions,
while the other two are held stationary. Since, as explained above, the center of mass is
positioned front with respect to the flagella, the vehicle rotates on the side of the two
rotating propellers. The results of the simulations are presented in Fig. 18, where some
snapshots of the simulations are presented and compared with the results of the
experiments performed with the same specifications. The modeled and the experimental
prototype present very similar behavior until a complete 90◦ rotation is obtained. Going
on, while the modeled vehicle continues rotating with a trajectory resembling a circle, the
real vehicle presents an ellipsoid-like trajectory. This discrepancy is due to the fact that
the vehicle is slightly negatively buoyant (a perfect neutral buoyancy is very difficult to
obtain with the proposed design) and the experiments are run in a relatively small tank,

Dept. of E&CE, SVCE, Bengaluru Page 27

 Title of seminar 2021-2022

where the flagella brush up against the tank’s bottom. In a future research, an improved
version of the vehicle will be investigated and the experiments will be performed in a real
pool. This prototype represents a first example of one of the many solutions that could be
obtained employing the flagellar modules as propulsive source. Different designs and
results can be attained optimizing the position, the number, the material and the design of
the modules, possibly resulting in increased overall performances of the robot.

45

Chapter 5

CONCLUSION

By centering the above issues, comes the need to create and plan an ease water quality observing
framework that can screen water quality progressively utilizing IoT environment.

Dept. of E&CE, SVCE, Bengaluru Page 28

 Title of seminar 2021-2022

REFERENCES

[1].SGeetha, S Gouthami“Internet of things enabled real time water quality monitoring

system”,Smart Water,Vol.2, No.1,pp. 2-19, 2017.
[2].

Dept. of E&CE, SVCE, Bengaluru Page 29