LOCOMOTION SIMULATION AND SYSTEM INTEGRATION OF ROBOTIC … · 2013-06-15 · K.H. LOW: LOCOMOTION SIMULATION AND SYSTEM INTEGRATION… I.J of SIMULATION, Vol.7, No. 8 ISSN 1473-804x

K.H. LOW: LOCOMOTION SIMULATION AND SYSTEM INTEGRATION…

I.J of SIMULATION, Vol.7, No. 8 ISSN 1473-804x online, 1473-8031 print 64

LOCOMOTION SIMULATION AND SYSTEM INTEGRATION OF ROBOTIC FISH WITH MODULAR UNDULATING FIN

K. H. LOW

School of Mechanical and Aerospace Engineering

Nanyang Technological University Singapore 639798

[email protected] Abstract: This paper presents a design of an underwater vehicle mimicking undulating fins of fish. To mimic the actual flexible fin of real fish, a fin-like mechanism is modeled with a series of connecting linkages. By virtue of a specially designed strip, each link is able to turn and slide with respect to the adjacent link. The driving linkages are used to form a mechanical fin consisting of several fin segments, which are able to produce undulations, similar to those produced by the actual fin rays. Owing to the modular design and the flexible structure of the mechanical fin, we are able to model different types of fin and to construct various biomimetic robots of fish swimming by fin undulations. Workspace of the fin mechanism is derived and the parametric study of the locomotion is presented. The integrated mechatronics design of the fin mechanism together with a buoyancy tank is briefly described. Some qualitative and workspace observations by experiments with the robotic fish are also shown and discussed. Keywords: Biomimetic, Robotic fish, Modular mechanism, Undulating fin motion, Workspace and locomotion, Buoyancy tank.

1. INTRODUCTION

Novel propulsion ideas have emerged recently as progress in robotics, new materials, and actuators have become available. Propulsors and vehicles that emulate the motion of fish have been developed for technological application, using state-of-the-art technology, robotics and an understanding of how fish swim. On the other hand, fin-based propulsion systems perform well for both high-speed cruising and high maneuverability in fishes, making them good models for propulsors of underwater vehicles.

Biomimetics is an emerging field, using

principles from living organisms to derive man-made mechanisms and machines that are capable of emulating the performance of animals [1]. In the field of underwater research, undulating-finned robot offers exceptional advantage over propeller in preserving an undisturbed condition of its surroundings for data acquisition and task execution. Military and defence are most important areas where biomimetic finds its significant role in ensuring safe waters; the undulating fin robot might be undetected when swim with a school of fish and therefore may act as a spy.

In hydrodynamics, highly accurate robotic

mechanisms are used to measure the power needed and forces generated, as well as to visualize the flow around their external skin structure [2]. Kato [3] and Read [4] studied the use of fins in small underwater vehicles for hovering and low-speed maneuvering.

Barrett et al. [5] used a laboratory robot to study

the effect of a fishlike body motion on the axial and transverse forces and hence the power required for swimming. The development of their high-precision multi-link robotic mechanism, which can emulate very closely the swimming of the tuna [6-8], overcomes the difficulties associated with working with live fish because it allows the acquisition of detailed measurements of the forces on an actively controlled flexible body.

As fish are impressive swimmers in many ways,

it is hoped that submersible robots that swim like fish might be superior to submersibles using propellers. Fish-like robots are expected to be quieter, more maneuverable and possibly more energy-efficient. In addition, each species has its own unique and optimum way of interacting with its environment, which then dictates the species’ body shape, body size, and the way it propels itself. This leads to the biological study and biomimetics development and study of robotic fish by many researchers [9-13].

To focus mainly on the biomimetic design of actual flexible fins, a robotic fish with a specially designed modular fin mechanism and buoyancy body is constructed. The fin mechanism provides the function of flexible membrane by using a slider. The modular concept enables us to construct various biomimetic robots of fish swimming by fin



undulations. The locomotion implementation and buoyancy control of the robotic fish are also presented and discussed. Finally, the preliminary testing of the robotic fish is discussed with the results presented.

Figure 1: Fish swimming with undulating fins.

2. DESIGN OF MODULAR FIN MECHANISMS 2.1 Design of Fin Rays

Breder [14] proposed two swimming modes of fish based on the propulsive structure used: body and/or caudal fin (BCF) locomotion and median and/or paired fin (MPF) locomotion. The present study focuses mainly in MPF fish, whose fins undulate roughly parallel to the direction of motion, as shown in Figure 1. Thrust is produced by large undulations along pectoral fins, which span from the anterior to the posterior of the fish. The amplitude envelope of the undulations increases from the anterior part to the fin apex and decreases toward the posterior.

Kier and Thompson [15] suggest that fins of

stingray are supported by three-dimensional array of muscle. The designs of actuators, both linear and rotary are unable to model the complex musculature of the fins of the stingray.

In order to simplify our modelling, the long fin

of fish is divided into many segments such that the fin looks similar to that of ray-finned fish. From the mechanical point of view, the fins of stingray can also be supported by rigid structure such as fin rays

as far as to perform undulatory movements. Figure 2 shows a fin diagram of any fish, including that of stingray, performing undulations. Universal joint (U) is one possible joint that permits two degree of freedom movements at the base of each fin ray.

Figure 2: Fin diagram of ray-finned fish.

2.2 Design of Flexible Membrane

At the present stage, the mechanical modelling of undulating fins, which includes those of stingray and ray-finned fish, is simplified to one degree of freedom from originally two degrees of freedom at the base of each fin ray. As shown in Figure 3, a servomotor serves as a muscle producing one degree of freedom at the base of each ray. A crank is next attached to the servomotor to function as a fin ray. In order for the fin to exhibit undulations similar to that of any undulating fin, each servomotor is programmed so that the crank attached to the servomotor oscillates based on a specified sinusoidal function.

The developed mechanism will enable the

flexible membrane to maintain a straight line between two cranks and at the same time extendible, since the distance between two cranks changes as they oscillate. Figure 3(b) shows the kinematic diagram of a linkage representing two cranks (AB and ED) and a membrane BD. The linkage in Figure 3(b) possesses two degrees of freedom since points A and E are each driven by a servomotor. Details of the designed mechanisms can be found in [16-18].



2.3 Workspace of Fin Segment

The analysis of the fin is simplified to one fin segment, which is controlled by two servomotors A and E, whose angular displacements with respect to the horizontal axis are specified respectively by θ1 and θ2, as shown in Figure 3. Note that the linkage in the figure has two degrees of freedom and BD is a link whose length varies according to the orientations of cranks AB and DE.

The position vector and the length of link BD are

respectively

jijiPPP

2211 sin)cos/()sin(cos θθθθ ++−+=−=

lsllDBBD

(1)

[ ] [ ]2122

12 )sin(sin)cos(cos θθθθ −+−+= llsBD and its orientation with respect to the horizontal line is

⎟⎟⎠

⎞⎜⎜⎝

⎛−+

−= −

12

121coscos/

sinsintan

θθθθ

φls

(2) Consider the fin segment in Figure 3 moving at a specified speed and the two servomotors performing harmonic oscillations according to the following functions:

)sin( and sin 21 βαθαθ +== ba (3) where a and b are the maximum angular oscillations of θ1 and θ2, respectively, while β is the phase difference between θ1 and θ2.

Spine shown as

fixed frame 1Servomotorsshown aspoints A & E

Cranks 2 & 5

Fin rays or spokesshown as revolute

joints B & DSlider C

Flexible membranemade of thin acrylicsheet, shown aslinks 3 & 4

Finrays

to beinserted

(a)

(b) Figure 3: A segment of fin pair. (a) CAD model. (b) Kinematic diagram.

An important advantage possessed by the mechanism shown in Figure 3 is that it will perform a predictable shape. Points B and D will always join with a straight line. The mechanism design allows the flexible membrane, represented by links 3 and 4 and the slider C, to contract to a minimum length of 80 mm and to be extended up to 130 mm. This therefore poses a certain constraint in the workspace of the mechanism, as shown by the gray area in Figure 4. There are two scenarios that will happen due to the mechanical limitation: (1) The slider C will hit end of track when the mechanism tries to contract to a length less than 80 mm, and (2) The slider C will run off track when the mechanism tries to extend beyond 130 mm. Scenarios 1 and 2 in Figure 4 are then the boundaries of the workspace, which are shown as gray line and dashed-black line, respectively. For example, the length is of BD is 90 mm, between the minimum and maximum ranges, if θ1 = θ2 = 45o.

2.4 Workspace Limitation

After the desired amplitude envelope and phase difference have been determined, it is important to check whether the undulating fin mechanisms will encounter any mechanical limitations in workspace. For the mechanisms to work properly, the angular positions of all fin segments must lay in the workspace defined in Figure 4. By using the same θ1-θ2 plot, Figures 5 and 6 show the trajectories traced at various β, which fall within the limiting ranges.



Dashed boundary: Slider C out of limitSolid boundary: Slider Chits end of track

-90

0

90

-90 0 90

1

23

6

5

4

θ1

θ290

-90

-90

90

0

2

3

4

B

C

θ1= 0°

θ2= -70°1

A E

D

1

2

3 4

5

A

BC D

Eθ1= 45° θ2= 45°

Scenario 41

2

34

5

A

BC

D

Eθ1= -60° θ2= -70°1

2

3

45

A

B

CD

Eθ1= 10°θ2= 80°

Scenario 3

Scenario 5

θ1= 70°θ2= -40°

1

2

3

45

A

B

C

D

Eθ1= -70°

θ2= 60°Scenario 6

1

23

4

5

A

B

C

D

E

Scenario 1

Scenario 2

Within workspace

Edge of workspace

Out of workspace

BD = 90 mm

BD = 97 mm

BD = 82 mm

BD = 75 mm

BD = 206 mm

BD = 209 mm

Figure 4: Workspace of the single fin segment due to any two adjacent motors, shown by gray area and black boundaries, which signifies the slider limitations (BD: minimum 80 mm, maximum 130 mm).

-90

0

90

-90 0 90

θ1

θ290

-90

-90

90

0

Trajectory ofangular positions ofservomotor whenundulating

Dashed-black boundary Slider off trackGray boundary Slider hits end of track

Figure 5: Workspace of undulating fin mechanisms of constant amplitude envelope and 45° phase difference β. All fin segments trace the same trajectory due to constant amplitude envelope. Note that the positions of servomotors refer to points B and D in Figure 3.

-90

0

90

-90 0 90

θ1

θ290

-90

-90

90

0

(a)

β = 0°

-90

0

90

-90 0 90

θ1

θ290

-90

-90

90

0

(b)

β = 10°

-90

0

90

-90 0 90

θ1

θ290

-90

-90

90

0

(c)

β = 60°

-90

0

90

-90 0 90

θ1

θ290

-90

-90

90

0

(d)

β = 90°

Figure 6: Trajectories traced by undulating fin mechanisms at various β. (a) β = 0°. (b) β = 10°. (c) β = 60°. (d) β = 90°. The size of the trajectory is proportional to the size of amplitude envelope.



Figure 7: Positions traced by undulating fin at various α and β.

By virtue of Eq. (1), the workspace of the fin

segments to generate such positions, with different servomotors and times, is shown in Figure 7 with three β (0, 30, and 60 degree) and two α (0.5 and 1 Hz). It is obvious that the wave number (N) increases for a higher phase difference (β), given the same servomotor number. On the other hand, the oscillation wave of a given motor increases, if the motor rotates faster (or a higher α).

3. FIN TRAJECTORY PLANNING CONTROL

Based on the design methodology described in Section 2, a finned propulsion system has been designed and is able to model undulating fins of underwater creatures, such as stingray, knifefish, cuttlefish, and many others, in order to investigate

various swimming methods. The locomotion implementation of the undulating fin is discussed next.

The multi-segment fin can accommodate various amplitude envelopes, motor frequencies, and wavelengths. The following equation is used to fit our programming purpose:

θn(t) =

gnsin2παnt+(n−1)βn, (4)

whose parameters are provided in Figure 8. The value n in Eq. (4) is used to indicate the number of servomotors, and the equation can therefore be seen as a discrete model of an undulating fin by a specified number of servomotors. For simplicity, the values of αn (also βn) can be kept the same for all servomotors.

The amplitude envelope gn defines the oscillation path of the respective crank of the servomotors, which can be mapped into one circular path. Other amplitude envelopes, such as linearly increasing amplitude envelope, can be described as follows: g

1 = c, g

2 = 2g

1, g

3 = 3g

1..., g

n = ng

1,

(5) where c is a constant. Note that the amplitude envelope has n number of circular paths for different diameters depending on the specified oscillation amplitude of each crank.

Figure 8: Parameters designation of undulating fin.

Figure 9: Arbitrary sinusoidal waveforms generated by a series of cranks connected by straight lines (for example, link BD with changing length.) As shown in Figure 9, the multi-segment fin mechanism is able to model any sinusoidal waveform, discretely by a series of straight lines – one straight line joins two points on a sinusoidal curve with varying waves. Figure 10 illustrates that



a higher number of the waves (or cycles) can be generated by the increasing number of servomotors. Figure 11 depicts the fin profiles generated by different sets of cranks at t = 0.1 sec. It is obvious that smoother waves can be generated by having more servomotors.

Figure 10: Discrete model of sinusoidal waveforms developed by series of straight-lines joining two points in every (a) 90° (i.e. 5 motors) and (b) 45° (i.e. 10 motors).



Figure 11: Fin segment coordinates with different sets of motors (6, 8, 10, and 12) at t = 0.1 sec. All segments are within the allowable ranges of link BD with slider.

Figure 12 illustrates the relationship of the servomotor frequency and the phase differences. The results show that the change of frequency does not affect the wave number, but the initial joint value. On the other hand, the increase of phase difference produces a higher wave number. Figure 13 shows the servomotor angles at various times with different sets of motors. It is seen that the wave number

increases, if more motors are used.

(a)

(b)

Figure 12: Servomotor angles (a) with different rotations, α; from 0.4 to 2.8 Hz in an interval of 0.4 Hz (b) with different phase differences, β, from 0 to 90 degree in an interval of 15 degree.



Figure 13: Servomotor angles at various times, from 0 to 0.6 second in an interval of 0.1 second, with different sets of motors (6, 8, 10, and 12).

4. CONTROL OF UNDULATING FIN MOTION

As an example, a robotic knifefish with anal fin having eight servomotors shown in Figure 14 is controlled by six microcontrollers, divided in three levels of communication for efficient programming.

Undulating fin

Controlboard

Level 1, two microcontrollersas clock & data collectorLevel 2, two microcontrollerscalculating sinusoidal functionLevel 3, two microcontrollers outputcalculated data to servomotors; eachmicrocontroller serve a fin module

Potentiometers

Figure 14: Layout of fin control board consisting of six microcontrollers, arranged in three levels of communications, and three potentiometers that alter some parameters of undulating fin.

As shown in Figure 15, there are two microcontrollers in Level 1 acting as a data miner and a clock. The data miner is used to obtain user input command, and the clock is used to synchronize processes of all microcontrollers. A 25-millisecond clock time was specified by the servomotors as the maximum time to process one cycle of instructions.

The computations in Level 2 are done in 15 milliseconds, and the data is next transferred to a corresponding microcontroller in Level 3. The technique is the so called parallel processing because all microcontrollers are doing different tasks, which are all synchronized by a clock. The overall operation of the control architecture is summarized in Table 1.



P6

Serial input

BS2sx (4)Calculator BS2sx (3)

Calculator

BS2sx (6)Data miner - potentiometer

BS2sx (5)Clock generator(synchronizer )

Serial inputPosition servos

Position servos

P7

P0

P8

P9

BS2sx (1)Effector

Position servos

P7

P15

Serial input

Serial input

Serial input Serial input

P6

P7

P9 P10

P4

P5

P5

P6

P4

P5

Position servos

P2

P3

P2

P3

10 KΩ

10 KΩ

10 KΩ

10 KΩ

VDDP2

P3

P4

P0

P0

P7

P0

P8

P9

BS2sx (2)Effector

P14

P15

220 Ω

10 KΩ

0.22 µF 220 Ω

10 KΩ

0.22 µF

10 KΩ

220 Ω

10 KΩ

0.22 µF

Communication legend:Flag busData bus Clock busDiagnose bus(probe)

25 ms

Level 1

Level 3

Level 2

Level 3

Level 2

P14

Figure 15: Circuitry for the fin locomotion control. Table 1: Control architecture using parallel processing technique.

Clock Level 1 Data Acquisition

Level 2 Computation

Level 3 Control

BS2sx (5) (Figure 15)

BS2sx (6) BS2sx (3) and BS2sx (4)

BS2sx (1) and BS2sx (2)

t0 = 0 ms Collect data from potentiometer, and transfer data to Level 2.

Compute based on a predefined value, and transfer the result to Level 3.

Waiting for the result from Level 2.

t1 = 25 ms Collect new data from potentiometer, and transfer data to Level 2.

Receive data from Level 1, compute based on the value at t0, and transfer data to Level 3.

Control servomotors based on the result at t0.

t2 = 50 ms Collect new data from potentiometer, and transfer data to Level 2.

Receive data from Level 1, compute based on the value at t1, and transfer data to Level 3.

Control servomotors based on the result at t1.

…... …... …… …... tx = 25x ms Collect new data

from potentiometer, and transfer data to Level 2.

Receive data from Level 1, compute based on the value at tx-1, and transfer data to Level 3.

Control servomotors based on the result at tx-1.

5. BUOYANCY TANK

Archimedes’ Principle states that any body partially or completely submerged in a fluid is buoyed up by a force equal to the weight of the fluid displaced by the body. A submarine varies its weight by filling tanks with water so that it submerges, and it can adjust its depth by adding water or ejecting water with compressed air. As for fish, their gas bladder has flexible walls that contract or expand according to the surrounding pressure [19]. The bladder has a gas gland that can introduce oxygen to the bladder to increase its volume and thus increase buoyancy. To reduce buoyancy, gases are released from the bladder into the blood stream and then expelled into the water via the gills, as shown in Figure 16.

Figure 16: Natural buoyancy in fish by the gas bladder.

In order to produce a fully deployable undulating-finned robot by the principle of swim bladder, a two-piston variable density chamber as a buoyancy body is designed and attached to the undulating fin mechanism, as shown in Figure 17. The buoyancy tank applies the similar principle by extending or retracting the two pistons using lead screws. This will change the amount of fluid displaced and thus the buoyancy force acting on it. The two sides of the buoyancy tube act as trim tanks. By adjusting the pistons, the amount of water in each side changes to maintain a horizontal angle. For example, if the right side of the buoyancy tank is higher than the left side, the piston on the right side will retract and this will bring the buoyancy tank to a horizontal angle. In the case that the right piston is fully retracted, the left piston will extend and again bring the buoyancy tube to a horizontal angle.



Figure 17: A two-piston variable density buoyancy tank to be attached to the fin mechanism.

5.1 Electronic Configuration and Program for Buoyancy Control

The simplified signal mapping for the buoyancy is shown in Figure 18. Note that the electronic configuration for the undulating fin (see Figure 15) is separated from the electronic configuration of the buoyancy tank with a separate power source. To demonstrate the depth-control implementation of the robotic fish, a DepthController program flowchart is shown in Figure 19. The program first obtains the actual depth and tilt angle from the depth and pitch sensors. The actual depth is then compared with the desired interval of allowable depth. The decision to extend or retract depends o the output signals from the four Opto proximity sensors, which help to determine if any of the pistons has reached either the inner or outer limits of the buoyancy tank.

Figure 18: Electronic configuration of buoyancy tank.

Figure 19: Depth_Controller program flowchart.



6. INITIAL EXPERIMENTS

Using the same design and control methodology, a robotic knifefish with anal fin has been constructed and tested. Figure 20 shows an experimental setup to capture and measure the movement of the designed knifefish robot, whose specifications are listed in Table 2.

(b)

Desiredswimming path

Robot’sinitial

position

70 cm80 cm

80 c

m40 c

m

240 cm

80 cm

40 cm

125

cm

40 cm

Video camera

Fixed ref. pointsRef. point

(a)

Buoyancy tank

Servomotors

Anal fin30 cm

x

z

y

Pitch

Yaw

Roll

xz

Figure 20: (a) Knifefish robot swimming, by anal fin, and submerged 30 cm underwater. (b) Top view of experimental setup in a 240 cm by 80 cm water tank and a video camera.

Table 2: Outline of knifefish robot specifications.

An important factor that must be taken into

consideration when capturing an object moving at a

certain frequency is the capture rate or sampling frequency. In the experiments, a video camera with sampling frequency of 25 frames per second was used to capture the robot’s motion, whose undulating

frequency is set for 98 Hz.

Note that an undulating fin is normally characterized by the number of waves (or cycles) N presented, which is related to β (in radian) by N = βnf/(2π), (6) where fn is the number of fin segments connected by (nf + 1) servomotors (nf = 7 for robotic knifefish modelled here).

Figures 21 and 22 show the velocity measured and generated for eight experiments with different β. The four experimental results in Figure 21 show a pattern with periodical variations. Another note-worthy observation of Figure 21 is that the pattern repeats itself at a frequency near that of the

undulating frequency, which is 98 Hz.

Roll and yaw oscillations (see Figure 20(a) for

definition) were not apparent by direct observations in experiments 5, 6, 7, and 8 in Figure 22. The velocity curves obtained from these experiments reveals roll and yaw were completely absent when the robot is moving forward. Their absence seems to confirm the analysis made by Sfakiotakis, Lane, and Davies [20], and Lindsey [21] stating that at least one wave (or cycle) is present when a fish is swimming.

Figure 22 also shows relatively constant velocity

curves obtain from experiments 5, 6, 7, and 8. Experiments 7 and 8 show small oscillations, as the robot was gaining speed from rest. Relatively constant velocity was observed after 1 second. In other words, a constant velocity and the absence of roll and yaw are desirable since the robot’s swimming path can be fully controlled. Experiments 5, 6, 7, and 8 also showed that the robot could move in a straight line, when more than one wave (N > 1) was present in undulations.

Biomimetic knifefish robot

Undulating fin

Mass 6.2 kg Mass 3 kg

Length 80 cm Length 63 cm

Height 56 cm Material Acrylic

Width 11 cm Width 20 cm

Actuator 8 servomotors, Futaba S3801

Peak-to peak amplitude

10 cm

Power source

7.5 V, 3000mAh NiMHbattery

Controller 4 parallel processing Basic Stamps

Fin segments

7 segments (8 servomotors)



Figure 21: Velocity of various swimming

modes for N < 1 (α = 98 Hz): (a)

Experiment 1 (β = 20°), (b) Experiment 2 (β = 30°), (c) Experiment 3 (β = 40°), (d) Experiment 4 (β = 50°). The patterns display periodical variations.

(a) (b)

(d)(c)

Velo

city

(cm

/s)

Time (s) Time (s)

Velo

city

(cm

/s)

Time (s) Time (s)

Velo

city

(cm

/s)

Velo

city

(cm

/s)

Figure 22: Velocity curves when N >1 (α =

98 Hz): (a) Experiment 5 (β = 60°),

(b) Experiment 6 (β = 70°), (c) Experiment 7 (β = 80°), (d) Experiment 8 (β = 90°). Nearly constant velocity was observed in these experiments.

The number of waves does not seem

proportional to the increase of velocity, as shown in Figure 23. In the white area (N > 1), the velocity is declining, as more waves were present in undulations. The gray area shown in the figure, in which periodical velocity variations are observed, is not of interest as it only contributes to inefficient swimming [20, 21]. Note that the darker area (smaller N) suggests more apparent periodic velocity variation, and roll and yaw oscillations. Figure 23 also depicts that the maximum velocity occurs near β

= 50o, the running frequency of the motor. At this speed, no roll, yaw and pitch were observed. With the absence of roll, yaw and pitch, more of the input energy will be transforming to energy to drive the knifefish robot forward. Other experimental results and discussion can be found in [22].

Velo

city

(cm

/s)

0

25

0 90

Number of waves, N1/ 2 3 / 2 7/ 41

β = 20°

β = 30°

β = 40°

β = 50° β = 60°β = 70°

β = 80°

β = 90°

N=7β / 360°

Figure 23: Velocity of 63-cm fin obtained from

water tank experiment (α = 98 Hz). Due to the

oscillatory nature of velocity for β less than 60°, the velocity in the gray area is plotted based its time-averaged value.

7. CONCLUSION AND FUTURE WORKS

An integrated robotic fish has been developed together with the fin rays, which is able to oscillate in arbitrary waves at a certain phase lead or lag. The integrated system consists of a buoyancy fish body attached with a set of flexible fin. The designed modular fin mechanism can also be re-designed to better mimic the swimming modes of fish with different fin types. Arbitrary or non-hormonic fin waves can be achieved by varying the rotation and phase difference of the servomotors. Workspace and locomotion of the fin mechanism have been studied with respect to various parameters.

More thorough study on fin locomotion and

hydrodynamics should be done in order to render proper and efficient control action on the current robot. An efficient force measurement of the robotic fish will be useful to the locomotion study. Therefore, future experiments should focus on how to obtain the force and the power of the undulating fins. The results enable us to determine the effective undulations for energy saving.

Future focus will also be on data acquisition from experiments, which can then be used to analyze the deployability – feasibility for deployment – of the robotic fish. Useful areas can also include the



mission execution by a school of robotic fish lead by a master or parent unit. The communication and coordination among the units are especially crucial in these applications. Research methodology and technology of vehicle’s navigation is also crucial to the localization and the recovery of robot fish in open water. It will be worthwhile to explore the use of sensors that could imitate the lateral line sensory organs of fish that they use to regulate their swimming in response to external “disturbances”.

ACKNOWLEDGEMENTS The author wish to thank Dr. F. M. J. Nickols for his valuable assistance in the early stage of the development of the first prototype. Helps given by Miss T. S. Indrawati, Mr. A. Willy, and Mr. W. L. Aw on the prototype development are greatly appreciated. The author would also like to acknowledge the support from Dr. Gerald Seet, the Director of Robotics Research Center (RRC), and the technicians at the RRC on the water tank, electronic equipment, and other facilities.

REFERENCES 1. Fish F. E., “Limits of Nature and Advances of

Technology: What Does Biomimetics Have to Offer?”. In Proc. of 1st Int. Sym. on Aqua Bio-Mechanisms, N. Kato and Y. Suzuki, Eds., Tokai University Pacific Center, 3-12, Hawaii, 2000.

2. Hover F. S. and Triantafyllou M. S., “Some

Robotic Applications in Fluid Mechanics: Vortex-Induced Vibrations and Fish Propulsion”. In Proc. Am. Soc. Mech. Eng. Fluids Eng. Div. Summer Meeing, Washington, DC, 1998.

3. Kato N., “Locomotion by Mechanical Pectoral

Fins,” J. Mar. Sci. Technol. 3:113–121, 1998 4. Read D., Fins for control and maneuvering of

underwater vehicles. MS thesis, MIT, Cambridge, MA, 1999.

5. Barrett D. S., Triantafyllou M. S., Yue D. K. P.,

and Grosenbaugh M. A., Wolfgang M., “Drag Reduction in Fish-Like Locomotion”. In J. Fluid Mech. 392:183–212, 1999.

6. Barrett D. S., Propulsive efficiency of a flexible

hull underwater vehicle. PhD thesis, MIT, Cambridge, MA, 1996.

7. Triantafyllou M. S. and Triantafyllou G. S., “An Efficient Swimming Machine,” Sci. Am. 272:64–70, 1995.

8. Wolfgang M., Tolkoff S., Techet A., Barrett A.,

and Triantafyllou M. S., “Drag Reduction and Turbulence Control in Swimming Fish-Like Bodies”. In Proc. Int. Symp. Seawater Drag Reduction, Newport, RI, Pp. 463–69 July 1998.

9. McHenry Matthew J., Pell Charles A., and Long

John H. Jr., “Mechanical Control of Swimming Speed: Stiffness and Axial Wave Form in Undulating Fish Models”. In The Journal of Experimental Biology 198, Pp. 2293–2305, 1995.

10. Rosenberger Lisa J. and Westneat Mark W.,

“Functional Morphology of Undulatory Pectoral Fin Locomotion in The Stingray Taeniura Lymma (Chondrichthyes: Dasyatidae)”. In The Journal of Experimental Biology 202, Pp. 3523–3539, 1999.

11. Lauder G. V. and Drucker E. G., “Morphology

and Experimental Hydrodynamics of Fish Fin Control Surfaces.” In IEEE J. Oceanic Eng. 29(3), Pp. 556-571, July 2004.

12. Westneat Mark W., Thorsen Dean H., Walker

Jeffrey A., and Hale Melina E., “Structure, function, and neural control of pectoral fins in fishes,” IEEE J. Oceanic Eng. 29(3), 674-683, July 2004.

13. Standen E. M. and Lauder G. V., “Dorsal and

Anal Fin Function in Bluegill Sunfish Lepomis Macrochirus: Three-Dimensional Kinematics During Propulsion and Maneuvering.” In The Journal of Experimental Biology 208, 2753-2763, 2005.

14. Breder C. M., “The Locomotion of Fishes.” In

Zoologica 4, 159-297, 1926. 15. Kier W. M. and Thompson J. T., “Muscle

Arrangement, Function and Specialization in Recent Coleoids.” In Proc. Coleoid Caphalopods through Time, Berlin, September 2002.

16. Willy A. and Low K. H., “Development and

Initial Experiment of Modular Undulating Fin for Untethered Biorobotic AUVs.” In Proc. of the Int. IEEE Conf. Robotics and Biomimetics (ROBIO05), Hong Kong, 2005.

17. Willy A., Design and development of undulating

fin, Master Thesis, School of Mechanical and



Aerospace Engineering, Nanyang Technological University, Singapore, 2005.

18. Willy A. and Low K. H., “Initial Experimental

Investigation of Undulating Fin”. In Proc. of the Int. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS2005), Edmonton, Canada, 2005.

19. Helfman G. S., Collette B. B., and Facey D. E.,

“The Diversity of Fishes,” Blackwell Science. 20. Sfakiotakis M., Lane D. M., and Davies J. B. C.,

“Review of Fish Swimming Modes for Aquatic Locomotion”. In IEEE J. Oceanic Eng. 24, Pp. 237-252, 1999.

21. Lindsey C. C., “Form, Function and Locomotory

Habits in Fish”. In Fish Physiology VII Locomotion, Hoar W. S. and Randall D. J., Eds. New York: Academic, pp. 1-100, 1978.

22. K. H. Low, “Biomimetic Motion Planning of an

Undulating Robotic Fish Fin”. In Journal of Vibration and Control, 2006 (in press).

BIOGRAPHY

Kin Huat Low (K. H. Low) obtained his MSc and PhD degrees in Mechanical Engineering from the University of Waterloo, Canada, in 1983 and 1986, respectively. After spending two years as a Post-doctoral Fellow at the University of Waterloo, he joined the School of Mechanical and Aerospace Engineering of the Nanyang Technological University. His teaching and research interests include robotics, vibrations, impacts, machines, mechanisms, exoskeleton systems, and mechatronics design. He is the author of five books and more than 100 journal papers in the areas of biomimetics, robotics, rehabilitation, impacts, power transmission systems, structural dynamics and vibrations. He has been the founding member and the first Chairman of the IFToMM in Singapore. He is currently the Co-Chairman of the IEEE Robotics and Automation Society (RAS) Technical Committee on Manufacturing Automation. He is a member of IEEE, IASTED, SIAA and SiCToMM. He is the Editorial Advisory Board Member of International Journal of Advanced Robotic Systems. He is also a member of several technical committees, conference committees, and journal editorial boards. He has won awards at the two IEEE conferences on the research work in robotics fish and assistive exoskeleton, respectively.

Documents

LOCOMOTION SIMULATION AND SYSTEM INTEGRATION OF ROBOTIC … · 2013-06-15 · K.H. LOW: LOCOMOTION SIMULATION AND SYSTEM INTEGRATION… I.J of SIMULATION, Vol.7, No. 8 ISSN 1473-804x