BIOMIMETIC CONTROL ALGORITHM FOR THE BALANCE

AND LOCOMOTION OF WALKING SYSTEMS

Nicu George Bîzdoacă, Anca Petrişor

University of Craiova, Faculty of Automation, Computers and Electronics, Caiova, Romania

Hani Hamdan

SUPELEC, Department of Signal Processing and Electronic Systems, Gif-sur-Yvette, France

Khalid Al Mutib

King Saud University, College of Computer and Information Sciences, Riyadh, Kingdom of Saudi Arabia

Keywords: Robotics, Biomimetic approach, Control strategy, Stable state transition.

Abstract: Implementation of active control has much potential to contribute to the creation and construction of

innovative structures. This paper summarizes recent research of the authors that is the study of biomimetic

control solutions regarding balance and locomotion of robotic systems. A first goal of the work consists in

identifying solutions necessary to balance the individual systems. Research has been focused both on

systems with a single foot, but also on biped, tripods, quadrupeds, hexapods and octopods. Static balance is

achieved by a proper mechanical design (Bizdoaca and Petrisor, 2009), but also by a corresponding load /

tensioning actuators systems that can compensate for inertial elements that can lead to system stability limit.

Theoretical studies have been focused on developing an efficient stepping algorithm in environment with

strong uncertainties, known as SSTA algorithm. The article present a series of experiments made with servo

actuated and smart actuated (based on shape memory alloy, especially) walking biomimetic structures.

1 INTRODUCTION

Biomimetics (or bionics, biognosis, etc.) is an

abstract of “good design from nature”. Roughly

speaking, biomimetics is the concept of taking ideas

from nature and implementing them in another

technology.This concept is actually very old, for

example, the Chinese wanted to make artificial silk

3,000 years ago. Some biomimetic processes have

been in use for years. An example is the artificial

synthesis of certain vitamins and antibiotics. More

recently, the biomimetic concepts, ideas and

applications are increasingly reported.

For example, the latest new biomimetic study

reported in the journal Nature, according to the

current picks of biomimetic issues, is actually from

studying how ants avoid traffic jams, which has

numerous implications for many scientists to rely on

the behavior of ants or other natural systems to give

them clues as to how to design computer systems

that avoid overcrowded networks. Another

biomimetic example, as commented by Philip Ball

in the 26 February 2004 issue of Nature in (

Hong,

and Bruce, 2004), is on the use of microbes in waste-

water that could make a handy household battery.

In a more general setting, according to

(Whatis.com), biomimetic refers to human-made

processes, substances, devices, or systems that

imitate nature. The art and science of designing and

building biomimetic apparatus is called biomimetics,

and is of special interest to researchers in robotics,

artificial intelligence (AI), nanotechnology, the

medical industry, and the military. Other possible

applications of biomimetics include nanorobot

antibodies that seek and destroy disease-causing

bacteria, artificial organs, artificial arms, legs, hands,

and feet, and various electronic devices. One of the

more intriguing ideas is the so-called biochip, a

microprocessor that grows from a starter crystal in

295

George Bîzdoac

a N., Petri¸sor A., Hamdan H. and Al Mutib K. (2010).

BIOMIMETIC CONTROL ALGORITHM FOR THE BALANCE AND LOCOMOTION OF WALKING SYSTEMS.

In Proceedings of the 7th International Conference on Informatics in Control, Automation and Robotics, pages 295-300

DOI: 10.5220/0002951102950300

 SciTePress

much the same way that a seed grows into a tree, or

a fertilized egg grows into an embryo.

Biomimetics is now not at the stage in

generating new concepts and ideas because the

mother nature has already provided numerous

models for us to imitate. The key is the

implementation and development which is gathering

momentum only recently because the science base

can cope with the advanced techniques in various

areas such as biology, materials, electronics,

computing, communication and control etc. The idea

of extrapolating designs from nature and copying

them has entered into many areas of applied science,

most notably the synthesis of new materials. So, it is

no surprise that people tend to regard the

biomimetics as an interdisciplinary field in materials

science, engineering, and biology.

2 STUDY OF BIOMIMETIC

CONTROL SOLUTIONS BASED

ON THE BALANCE OF

INDIVIDUAL SYSTEMS

Studies of the authors have been focused to identify

the necessary solutions to ensure the individual

systems balance. The research have been directed

both on systems with a single foot, which made a

movement by jumping, but also on biped, tripod,

quadruped, hexapod and octopod systems.

Static balance is achieved by a proper

mechanical design [CMSM'2009], but also by a

corresponding load / tensioning of actuators systems

that can compensate the inertial elements that can

lead the system to the stability limit.

In addition to these studies for extension degrees

of freedom in order to provide a more efficient

navigation and balance control have been explored

and studies smart damping and actuation based not

only on servoactuators, but and to shape memory

materials and magnetorheological fluids.

The essential condition for static equilibrium

related to the position of the system gravity centre of

which projection must be inside the contact surface

described by the support elements (feet) of the

biomimetic system is, for jumping and biped

systems extremely difficult, while for ensuring a

dynamic balance the use of specialized control

architectures is required.

Figure 1: Study of biped locomotion.

Taking into account the human model, it is

observed that it has a balanced structure in addition

to appropriate state, and a sensorial system that

provides the driving of the muscular system to

compensate for factors that could lead to system

stability limit.

Therefore, to achieve the balance of studied

systems, it turned to a series of sensorial elements to

ensure the system proximity identification of the

stability limit, while an architectural simplification

of the system to ensure absolutely necessary degrees

of freedom for a dynamic compensation but much

more limited than the biological model.

These considerations are related to the

complexity of calculations necessary to some

systems with more degrees of freedom, respectively

considerations of ensuring the bionic system

mobility, especially with a lower energy

consumption.

Calibration bionic system proved to be a basic

element, that essentially influenced subsequent

locomotion of the mechatronic system.

Figure 2: Biped biomimetic structure.

An important component of the study was

focused on walking structures issues - quadruped,

hexapod and octopod structures.

ICINCO 2010 - 7th International Conference on Informatics in Control, Automation and Robotics

296

Figure 3: Octopod biomimetic structure.

Figure 4: Quatrupede biomimetic structure.

Consulting literature revealed some aspects of

the modalities and criteria for classification of

walking robots structures, thus achieving a hierarchy

in terms of biological models imitated, of physical

morphology, of functions they performed and of

locomotion strategy that they adopt.

In order to have a more complete image of the

themes (walking robots field), we focused on:

• walking stability issues of such robots

• design advantages of such morphologies

• issues of wide range of areas where they

prove their aplicability.

Conclusions resulted in this work led to the need

for a dual approach: intelligent design and balancing

of the mechanical system and implementing of a

system/algorithm that allows the achievement of the

instability system compensation or its approach to

the stability limit, by the dynamic analysis of the

specific conditions.

The developed algorithms should be simple,

robust, adaptive, sub-optimal for reasons related to

the response time.

3 STUDY OF BIOMIMETIC

CONTROL SOLUTIONS BASED

ON THE LOCOMOTION OF

INDIVIDUAL SYSTEMS

Individual systems locomotion behave correctly

identification of the control algorithm, but also, the

identification of some kinematic and dynamic

models, viable and efficient in implementing of

these algorithms through command and control

architecture.

Studies of systems moving by jumping with a

single leg, with 4 feet, respectively hiperredundante

systems, systems that have shape memory springs in

their structure, have revealed the effectiveness of

this type of solution for continuous monitoring of

the damping coefficient necessary at the contact with

the ground or for damping and driving of the

mechatronic architecture (Bizdoaca, 2009).

Figure 5: Mathematical model of a one-legged robot that

uses as damping element a shape memory spring.

Figure 6: Biomimetic structure type rabbit.

Experimental measurements in conjunction with

theoretical and practical work of the research team,

have enabled studies to be efficient.

BIOMIMETIC CONTROL ALGORITHM FOR THE BALANCE AND LOCOMOTION OF WALKING SYSTEMS

297

Figure 7: Biomimetic structure type frog. DTA and DSC

curves for 18.275mg from SMA spring.

The effect of the energesing smart damping

system is extremely efficient, numerical simulations

showing this.

Figure 8: Hopping robot trajectory for different energizing

temperatures of the SMA spring.

Particular attention was allocated for the

hiperredundante systems both regarding creeping

locomotion and handling structures as trunk type.

Kinematic study of walking systems was

performed both on the robot body movement in

space, given the motion laws of a rigid body in

three-dimensional space and on the robot foot

considering that it has a structure consisting of three

links and three degrees of freedom in relation to the

body, achieved by three rotational joints.

The rotation axes for the two joints which form

the hip are set to intersect orthogonally.

In the case of existing a defect of the bionic

architecture, locomotion was a complement of the

activity of the project.

In robotic structures, the faults may be caused by

external environmental conditions (operating

environment) or by internal conditions (structure,

sensors, actuators or control).

Faults monitoring system, according to its

complexity, may adopt the following control

strategies (from simple to complex): only detection

and location of faults and possible suggesting of

isolation actions and/or avoiding of fault

components; further robot movement until bringing

it into a safety state in a neighborhood of the end

point for major faults; the robot stop in a safety state

with maintaining of stability in the case of

catastrophic faults.

General techniques for detection and

identification of existing faults and those for

recovery after failure, can be applied to walking

robots systems, thus determining the operation space

under fault. For this purpose were determined the

operation areas of the walking robots feet as an

annulus sector. Overcoming these allowed operation

areas for each leg lead to interference problems.

To avoid this, were a prori eliminated all the

areas that can be overcome, so that each leg has its

own separate region.

Thus, it was defined for each leg one operation

area (called operation cell) having a rectangular

shape. Then it was analized, from kinematic point of

view, the case of existing a fault due to a blocked

joint of a robot leg. This analysis of the fault can be

extended to other robot legs because of the robot

symmetry.

Based on kinematic restrictions mentioned

above, were presented and justified fault-tolerant

locomotion algorithms for each considered case of

fault both for the straight line robot moving and for

a crab stepping type: robot locomotion tolerant at

first joint blocking, robot locomotion tolerant at

second joint blocking, robot locomotion tolerant at

third joint blocking.

4 IMPLEMENTATION OF THE

WALKING ROBOT CONTROL

ALGORITHM IN SSTA

STRATEGY

It is considered the walking robot structure as

depicted in Fig.9, having three normal legs L

, L

and a head equivalent to another leg, L

, containing

the robot centre of gravity, G, placed in its foot. The

robot body RB is characterized by two position

vectors O

, O

and the leg joining points denoted R

, R

. The joining point of the head, L

, is the

ICINCO 2010 - 7th International Conference on Informatics in Control, Automation and Robotics

298

central point O

, R

= O0, so the robot body RB is

univocally characterized by the set,

01ijp0

RB {O ,O , , , , }=λλλλ

Figure 9: The geometrical structure of the robot.

A robot leg, let us consider of index i, has a body

joining point R

expressed by a complex number and

the foot point denoted by the complex number G

. It

contains two joint segments defined by the lengths

, b

with the angles α

, β

. Here there are considered

three legs: i, j, p.

Due to the complexity of the evolution in an

environment with strong uncertaines, an efficient

walking algorithm which ensure stability and

locomotion of the robotic structure, must be

developped.

Stable States Transition Approach (SSTA)

control strategy applied for a hexapod structure is

proposed. By SSTA strategy is assured the walking

robots evolution in uncertain environments

subordinated to two goals:

- achievement of the desired trajectory expressed

by the functions

Of(x)= and (x)θ=θ , where x is

the ground abscissa and

Ox= ; it is considered the

evolution from left to right;

- assuarance of the system stability that is, in any

moment of the evolution the centre of gravity has to

be in the stability area.

Considering the walking robot as a variable

causality dynamic system it is possible to realize this

desideratum in different variants of assurance the

steps succession. The steps succesion suposes a

series of elementary actions that are accomplished

only if the stability condition exists.

Continuously, by sensorial means or using the

passive leg, the robot has informations about its

capacity of evolving on the ground. Every time it is

considered that the legs i, j are on the ground and the

the system is stable (

[0,1]ε∈

). The passive leg

is which realises the walking.

By testing the ground is realized its division in

lots representing the fields on x axis which constitue

the abscissas of some points that can be touched by

the

G leg. A next support point given by the free

G leg , is chosen so that to existe a next stable state

or ε

, taking into account the actual state of legs

activity.

A variant of mouvements succesion, composed

by 12 steps, was proposed in (Petrisor, 2008).

5 EXPERIMENTAL RESULTS

An experimental platform, called RoPa, has been

conceived. The RoPa platform is a complex of

MATLAB programs for simulation and control of

walking robots evolving in uncertain environments

according to SSTA control strategy. A number of

eight causality orderings of the robot structure have

been implemented on RoPa.

Figure 10: RoPa Graphic User Interface.

-2 0 2 4 6

-4

-3

-2

-1

RoPa1Ex2S5V1q132cz15250

x [m]

y [m]

Figure 11: The robot kinematics evolution.

BIOMIMETIC CONTROL ALGORITHM FOR THE BALANCE AND LOCOMOTION OF WALKING SYSTEMS

299

The stability of this evolution is graphical

represented by a stability certificate of the evolution

(Figure 12).

This certificate attests the stability index of the

active pair of legs in any moment.

Figure 12: The stability certificate of the evolution.

6 CONCLUSIONS

Experiments on walking structures, made in the

mechatronics laboratory, revealed the efficiency of

SSTA algorithm, providing a robot system stability

especially on hard terrain or with a high degree of

uncertainty regarding the nature and topography of

the contact surface. Using smart materials in the

structure of biomimetic mechatronic architectures

lead to an extension of the control capabilities.

These latter elements will be explored in subsequent

theoretical developments of research activity,

practical experimentation, empirical, currently

performed being extremely encouraging.

ACKNOWLEDGEMENTS

This report is part of the Romanian National

University Research Council (CNCSIS) contribution

to the project PNCDI – II - 289/2008.

REFERENCES

Bizdoaca, N. G, Tarnita, D., Petrisor, A., Diaconu, I.,

Bizdoaca, E.,2009. Control and design for biomimetic

application using smart materials. Annals of DAAAM

for 2009 & Proceedings of the 20th International

DAAAM Symposium, "Intelligent Manufacturing &

Automation: Theory, Practice & Education", Volume

20, Nr.1, ISBN 978-3-901509-70-4, ISSN 1726-9679,

pp. 0949-0950, Published by DAAAM International,

Vienna, Austria.

Bizdoaca, N. G, Tarnita, D., Petrisor, A., Bizdoaca,

E.,2008. Modular Adaptive Bionics Structure, WSEAS

TRANSACTIONS on BIOLOGY and BIOMEDICINE,

Issue 9, Volume 5, ISSN: 1109-9518, pag 229-238

Centre for Biomimetics, “What is biomimetics, http://

www.rdg.ac.uk/Biomim/brochure.htm.

Cubero S., 2001. A 6-Legged Hybrid Walking and

Wheeled Vehicle. 7-th International Conference on

Mechatronics and Machine Vision in Practice, USA.

Fulgenzi, C., Spalanzani A., Laugier C., 2007. Dynamic

Obstacle Avoidance in uncertain environment

combining PVOs and Occupancy Grid, Robotics and

Automation, 2007, pp.1610-1616.

Galt, S., Luk L., 1999. Intelligent walking gait generation

for legged robots. Proc. 2-th International Conference

on Climbing and Walking Robots.

Hong, L., Ramanathan, R., Bruce, L., 2004. Production of

electricity during wastewater treatment using a single

chamber microbial fuel cells, Environ. Sci. Technol.,

vol. 38, no. 7. 10.1021/es034923g, pp. 2281 – 2285.

Jung, S., Hsia, T.C., Bonitz, R., 2004. Force Tracking

Impedance Control of Robot Manipulators Under

Unknown Environment. IEEE Transactions on

Control Systems Technology, 12(3), 474-483.

Kiguchi, K., Watanabe, K., Izumi, K., Fukuda, T., 2000.

Application of Multiple Fuzzy-Neuro Force

Controllers in an Unknown Environment Using

Genetic Algorithms, Proc. of IEEE International

Conference on Robotics and Automation.

Kiguchi, K., Miyaji, H., Watanabe, K., Izumi, K., Fukuda,

T.,2000. Design of Neuro Force Controllers for

General Environments Using Genetic Programming,'

Proc. of the Fourth Asian Fuzzy Systems Symposium

(AFSS2000), pp. 668--673.

Nanayakkara, T., Watanabe K., Izumi, K., Kiguchi, K.,

2001. Evolutionary Learning of a Fuzzy Behavior

Based Controller for a Nonholonomic Mobile Robot in

a Class of Dynamic Environments, Journal of

Intelligent and Robotic Systems, Vol. 32, No. 3.

Petrisor, A., Bizdoaca, N., 2008. Stable States Transition

Approach - a new strategy for walking robots control

in uncertain environments. ISI Proceedings of the 4th

International Conference on Informatics in Control,

Automation and Robotics, Madeira, Portugal.

Petrişor, A, Bîzdoacă, N., 2008. Implementation of the

Stable States Transition Control Algorithm for a four

free joints Walking Robot. WSEAS Transactions on

Systems Journal, Issue 1, Volume 8, ISSN 1109-2777,

pp.107-116.

Schmuacer U., Schneider A. Ihme T., 1996. Six Legged

Robot for Service Operations.

Proc. of EOROBOT’96;

IEEE Computer Society Press, pp:135-142.

Thirion, B., Thiry, L., 2001. Concurent Programming for

the Control of Hexapod Walking. 7-th International

Conference on Mechatronics and Machine Vision in

Practice, USA.

ICINCO 2010 - 7th International Conference on Informatics in Control, Automation and Robotics

300