Collective Energy Foraging
of Robot Swarms and Robot Organisms
Serge Kernbach
Institute of Parallel and Distributed Systems, University of Stuttgart
Universit¨atsstr. 38, D-70569 Stuttgart, Germany, Serge.Kernbach@ipvs.uni-stuttgart.de
Abstract— Cooperation and competition among stand-alone
swarm agents increase collective ﬁtness of the whole system.
A principally new kind of collective systems is demonstrated
by some bacteria and fungi, when they build symbiotic or-
ganisms. Symbiotic life forms emerge new functional and
self-developmental capabilities, which allow better survival of
swarm agents in different environments. In this paper we
consider energy foraging scenario for two robotic species,
swarm robots and symbiotic robot organism. It is indicated that
aggregation of microrobots into a robot organism can provide
better functional ﬁtness for the whole group. A prototype of
microrobots capable of autonomous aggregation and disaggre-
gation are shown.
I. INTRODUCTION
Natural collective systems demonstrate that many indi-
viduals cooperate when this is proﬁtable to each of them.
Examples are cooperative hunting of predatory animals,
group-based foraging of mammals [1] or nest building of
social insects [2]. In these and many other examples, animals
get together when it provides better chances for foraging,
for defence, or generally for surviving in their environment.
Participants of these groups can be weak with limited sen-
sors/actuator capabilities, however collectively they can build
a strong group with very extended capabilities.
Lately, technical systems mimic natural collective systems
in improving functionality of artiﬁcial swarm agents. Col-
lective, networked or swarm robotics are scientiﬁc domains,
dealing with a cooperation in robotics [3], [4]. Research in
collective robotics is mostly concentrated on stand-alone au-
tonomous robots. Cooperation and competition among stand-
alone robots increase their collective ﬁtness [5]. However,
natural swarm agents can build a principally new kind of col-
lective systems. For example, fungi dictyostelium discoideum
can aggregate into a multi-cellular symbiotic organism and
perform such activities that cannot be fulﬁlled alone or in
a swarm-like way [6]. The symbiotic organization emerges
new functional capabilities, which allow swarm agents to
achieve better ﬁtness in the environment [7]. When the need
of aggregation is over, symbiotic organism can disaggregate
and exists further as stand-alone agents [8].
Swarm robots can also build symbiotic life forms and
achieve better functional ﬁtness. To demonstrate this idea,
we consider a collective energy foraging scenario for mi-
crorobots Jasmine [9]. Swarm robots can autonomously
ﬁnd an energy source and recharge. Dedicated collective
strategies [10] can essentially improve the efﬁciency of
energy foraging, but nevertheless a functional ﬁtness of a
swarm is limited. For instance, if the recharging station is
separated from a working area by a small barrier, robots
can never reach the energy source. However, when robots
will aggregate into more complex symbiotic organism, which
can pass the barrier, they will reach the docking station.
In this paper we demonstrate main differences in hardware
and software as well as in behavioral strategies of both
robot systems: swarm robots and symbiotic organisms. Based
on the existing microrobotic platform Jasmine IIIp [11], a
prototype of multi-robotic organism is developed. A few
topological and functional issues of aggregation into an
organism are shown [12].
One of the main arguments for the collective energy
foraging is related to the idea of functional self-development:
robot organisms can autonomously change their functionally.
The energy homeostasis can appear as a task-unspeciﬁc cri-
terion, which deﬁnes such self-development. Since energetic
efﬁciency, by analogy to natural survival, possess open-
ended features, the collective energy homeostasis can be
considered as a candidate underlying a design of open-ended
self-developmental systems [13].
The rest of the paper is organized in the following way.
In Sec. II the energy foraging scenario and the correspond-
ing hardware and software of swarm robots are described.
Sec. III demonstrates collective strategies and limitations
imposed on collective ﬁtness of the robots. Sec. III brieﬂy
overviews the development of symbiotic microrobots. Fi-
nally, in Sec. V we conclude this work.
II. ENERGY FORAGING IN A ROBOT SWARM
The distinctive property of any living organisms is the
energy homeostasis and, closely connected, foraging be-
havior and strategies [1]. The robots, equipped with on-
board recharging electronics, can also possess its own energy
homeostasis [14]. In this way, when swarm robots become
”hungry”, they can collectively look for energy resources
and execute different strategies in cooperative energy for-
aging [15]. In critical cases robots can even decide to
perform individual foraging, competing with other robots for
resources. In this case we also see analogies to behavioral
strategies of animals [16].
The need of energy is a perfect example of natural ﬁtness.
When robots, with their strategies, found enough energy, they
survive in the environment. It means these strategies are good
arXiv:1111.0873v1  [cs.RO]  3 Nov 2011
enough, other case robots energetically die. Based on energy
foraging, several evolutionary approaches for different robot
species can be developed, compared and tested. Further in
the section we brieﬂy demonstrate the developed solutions
for swarm energy foraging and show a limit of functional
ﬁtness in a swarm.
A. RECHARGING HARDWARE
The hardware is described in
[15], [10] and generally
follows the idea of swarm embodiment [17]; here we brieﬂy
overview it. To make the microrobots capable of autonomous
recharging, four following components are required: (1)
internal energy sensors, monitoring energy level of Li-
Po accumulator; (2) especial recharging circuits for Li-Po
process; (3) reliable connectors to docking stations with a
low electrical resistance; (4) communication with docking
stations. Internal energy sensor is implemented as a resistive
voltage divider with the coefﬁcient 0.55, it consumes a
continuous current of 5µA. The voltage divider is connected
to a non-regulated power line and with an ADC of the
microcontroller.
The fully charged Lithium-Polymer battery provides the
micro-robot with energy for nearly 1,25 hours. Li-Po ac-
cumulators require speciﬁc recharging process. To control it
we use the LTC4054 Li-Ion/Li-Po battery charger. It is small,
with a low dropping voltage and can be directly installed on
the Li-Po accumulator. To perform autonomous recharging,
robots have to possess connectors to the docking station,
which allows simple docking and reliable mechanical contact
of low electrical resistance. After several tests [18] the best
solution was two 0.4mm silver-plated wires glued in the
front area of the robot. Both connectors are installed on
different height, see Fig.1(a). The docking station represents
(a)
(b)
Fig. 1. (a) Microrobot Jasmine IIIp, shown are recharging contacts;
(b) The docking station and a few robots recharging there. Images
are from [15].
a wall with glued 0.2mm thin copper stripes of 5mm wide,
see Fig. 1(b). Both copper stipes are connected to the 5V
source. Many such docking stations can be placed together,
see Fig. 2. To connect to the docking station, a robot has to
move to this wall (based on the docking signal), until it gets
a positive signal from the touch sensor. After that the robot
shortly turns on its wheels to produce a small mechanical
strain. The communication system of the docking station
sends continuously the signal ”I’m free slot in the docking
station” (coded numerically). This signal can be received
within 10-15cm away from the docking station. Discharged
robots, when sensing a free docking slot, approach the
docking station and start a recharging process. Body of the
recharging robot blocks the signal, so that no other robot can
receive the signal from this slot. When recharging is ﬁnished,
the robot moves away.
B. ENERGETIC HOMEOSTASIS OF SWARM ROBOTS
As described in the previous section, the robot is able
to sense its own energetic level and the position and avail-
ability of energy sources. Capacity of a single cell Li-Po
accumulator is enough for a running time of at least 1,25
hour. The optimal working mode of Li-Po accumulator is
discharging only up to 75-80% of capacity. The critical level
of accumulator is the voltage dropping less than 3V, because
in this case the internal power regulator is not able to stabilize
voltage ﬂuctuations and microcontroller can spontaneously
reboot. The full recharging takes about 90 minutes, the partial
recharging is almost equal to discharging (15 min. motion
requires about 15 min. recharging).
The energetic homeostasis of the robot is related to the
voltage level of the Li-Po accumulator and includes ﬁve
different states. When the voltage drops under 3.05V, a robot
should stop and go in stand-by mode. In this state it is
not able to react on external stimulus and needs a human
assistance for recharging. When the voltage is under 3.2V,
a robot should look for docking station independently of
the current task. It has about 3-5 minutes to ﬁnd it, other
case it will energetically die. When the voltage is less than
3.7V but more than 3.2V, the robot has different degrees
of discharging. It means a robot can start look for energy
sources, when there are no more important tasks. The lower
the energetic level is, the higher is the priority of looking for
energy. Ideally, when reaching 3.65V, robot should start look-
ing for the docking station. When during recharging a robot
achieves 4.0V, this state can be characterized as satisﬁed –
accumulator is not fully recharged (80-85%), but enough to
run again and make slot free for other robots. The voltage
increases between 4.1 and 4.2 very slow, (accumulator is
already recharged up to 90-95%). When voltage is about
4.2V (ADC value 196) the circuit stops recharging. In this
state the robot is fully recharged.
Firstly, in a critical state, robots should break the currently
executed collective or individual activity. This is not typical
in robotics, however a died robot can essentially distort
collectively executed activity. Secondly, a robot should have
the priority of a currently executed activity and the priority of
looking for energy sources. When, for example, the priority
of current activity is 0.6, but the level of discharging is 0.7,
robot will look for a docking station. Finally, a robot can
have so-called ”collective instinct”, it can recharge only till
”satisﬁed state” (it takes less time), and makes a slot free for
recharging of another robot.
Collective energy foraging can be designed on the one
hand as an artiﬁcial self-organizing process [19], which pos-
sess adaptive [20] or controllable-emergent [21] properties.
On the other hand, application of planning-based approaches,
e.g. [22], [23], is also possible. Generally it depends on
cognitive capabilities of the robot [24], in particular on the
sensing system [25]. Independently of the selected approach,
a cooperativeness of individual energetic homeostasis con-
sists in management of (a) a priority of collective tasks,
(b) robots with critical energy states or (c) the regarding
behavior.
III. COLLECTIVE ENERGY FORAGING AND
LIMITATION OF FUNCTIONAL FITNESS
The need of collective strategy in energy foraging appears
from the optimization of the swarm efﬁciency. The energetic
swarm efﬁciency Φ is calculated as a relation of the time
spend for useful activities tT ASK and the total time tT ASK+
tRECHARGING. Remembering that the working time is
equal to the recharging time, the best swarm efﬁciency is
Φ =
tT ASK
tRECHARGING + tT ASK
= 0.5.
(1)
This best efﬁciency is approximatively achievable when one
robot works alone (Φ = 0.45 −0.48). When there are many
robots, a few undesired effects can appear: (1) the docking
station can be a ”bottle neck” that essentially decreases the
swarm efﬁciency; (2) robots with a high-energy level can
occupy the docking station and block low-energetic robots.
These robots can energetically die (and so decrease the
swarm efﬁciency); (3) many robots can create a ”crowd”
around a docking station and essentially hinder a docking
approach. This can increase the total recharging time and
makes worse the energetic balance of the whole swarm.
Robots, in their energetic homeostasis, have only two
possible individual decisions: to execute a current collective
task or to move for recharging. In this way, a cooperative
strategy should ﬁnd a right timing and a right combination
between these individual decisions of all robots.
There are two ways to synchronize individual decisions.
Firstly, the decision making procedure can be done on the
collective level and all robots have only to execute the
obtained decision [26]. Examples of such a decision making
process are bargaining, auctioning or similar, where the ﬁnal
collective activity is not ﬁxed and should ﬁrst be negoti-
ated [27]. Application of numerical approaches for collective
decision making is also possible [28]. Secondly, the decision
can be done individually, however input information for this
decision will be prepared collectively. This decision making
process is usually applied for ”switching decision”, where the
ﬁnal collective activity represents a sequence of predeﬁned
sub-activities. Obviously, the individual state of a robot also
inﬂuences its own decision, so that ﬁnally a collective be-
havior represents a complex combination between collective
”needs” and individual ”desires”. Lately, several strategies of
energy foraging for a robot swarm up to 50 swarm agents are
implemented, see Fig. 2. These cover different bio-inspired
approaches [29], [30] and hand-coded strategies [31]. The
efﬁciency of these approaches is between 0.18 (mostly hand-
coded) and 0.33 (bio-inspired).
(a)
(b)
(c)
Fig. 2.
(a) Docking of a few robots for recharging. Shown is the
two-line approach: the ﬁrst line - recharging robots, the second line
- robots waiting for recharging; (b) The ”barrier problem” - robots
are separated form docking stations by a barrier; (c) A possible
solution to the ”barrier problem”: swarm robots form a multi-robot
organism and collectively pass the barrier.
In one of these experiments, a few robots died nearly the
docking station, so that they build a barrier from one border
to other one in the robot arena (we ”simulated” this in the
Fig. 2(b)). Robots located before this barrier (away from the
docking station) will energetically die. This is the limit of
a functional ﬁtness of swarm robots. There is no strategy
that allow swarm robots to overpass the barrier. Only when
swarm robots will collectively emerge new functionality,
such as ”pass the barrier”, they will solve the ”barrier
problem”. In the next section we consider this case.
IV. ROBOT ORGANISMS
An approach to solve the ”barrier problem” is an aggre-
gation into a multi-robot organism. They can achieve the
docking stations by ”growing legs” and overstepping the
barrier. So the robots are helping each other in a symbiotic
life form, see Fig. 2(c). Obviously, such a solution is ex-
tremely challenging from many viewpoints. Symbiotic robot
systems have many similarities with known robotic research
as e.g. mechanical self-assembling [32] or reconﬁgurable
robotics [33]: (1) Robots should be capable for autonomous
aggregation and disaggregation; (2) Robots in the disaggre-
gate state should possess individual locomotion; There is no
central control neither for disaggregated state (swarm) nor for
the aggregate state (organism), it should possess a kind of
guided self-organization [34]; (4) Stand-alone robots should
proﬁt from the aggregation into organism.
A. BRIEF OVERVIEW OF HARDWARE AND SOFTWARE
This section demonstrates a developed solution before
start of the projects [35], [36] within so-called internal pilot
projects. This version of robots is based on the Jasmine
III+ development, shown in Fig. 1(a). From the hardware
viewpoint, the most challenging part is to provide simple and
reliable docking mechanism, which allows an autonomous
aggregation and disaggregation. The successful solution is
shown in Fig. 3. The docking mechanism consists of one
male part, placed on the front of the robots and three female
parts, placed in wheels and on the back, see Fig. 3(c).
Male and female connectors can be locked with a hook,
as shown in Fig. 3(b). This mechanism is experimentally
tested in a small group of 13 prototype robots. Combining
different wheel- and back- connectors, various conﬁgurations
with many degrees of freedom (DOF) can be obtained,
see Fig. 3(d). Combining several DOF, the organism can
demonstrate different locomotion principles, as shown in
Fig. 3(e),(f).
The functional structure of the organism is shown in
Fig. 4(a). The organism should possess a common energy and
communication bus. Client electronics for both busses should
allow taking and giving energy from/into the bus and a high
speed serial data transfer. Special short-circuit protection
switches off such segments of energy and data busses that
are defect. In this way, even when some robots cause failure
on central busses, the organism will survive. Each robot
has 250mA/h Li-Po accumulator connected in parallel in the
organisms (with Li-Po balancer). When organism has about
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 3. (a) 3D Assembling diagram of the docking mechanism; (b)
Lock with a hook; (c) Implementation of the docking mechanism in
the prototype; (d) 5 degrees of freedom for the organism when using
two rotational connectors; (e),(f) Different locomotion principles,
obtained by combining several DOF;
30 robots, common energy resources are about 7,5A/h. With
even 4C short-time discharging, the organism can produce
about 30A current for application purposes (intended size of
the organism is about 30×30×30cm and wight about 500gr.).
When taking into account that only a few robots in the
organism perform actuation (and so have a maximal energy
consumption), the aggregation into the organism should be
proﬁtable for robots at least from the energetic viewpoint.
(a)
(b)
Fig. 4.
(a) The functional structure of the symbiotic organism; (b)
Computational structure of the symbiotic robot organism.
In the current development, each robot is controlled
through three Atmel microcontrollers (MCU), which commu-
nicate via internal I2C bus: MPU with external memory for
behavioral programs, MPU for control of sensors/actuators
and MPU for ZigBee and IR communication. Computational
resources of the organism are separated on three layers:
physical, middleware and virtual genomic one, see Fig. 4(b).
MPUs communicate via data bus and build a distributed com-
putational system. Each robot supports middleware, which
allows running user-deﬁned tasks on many MPUs in parallel
and uses distributed memory resources. The middleware sys-
tem takes technical care about introducing new members into
the organism or excluding some robots from the organism.
The main MPU is programmed by so-called genetic frame-
work [37], [13], [38]. A genome of an artiﬁcial organism
carries the total set of genes and includes instructions for
building, running and maintaining the organism. Genes are
functionally complete and consist of small ’states’ (by anal-
ogy to biological ’bases’ [39]). For example, the function
’move’ is a gene and consists of different states that control
a movement. Using these states, it is possible for genes as
well as for the whole genomes to be manipulated through
recombination or mutation. Genome is stored in external
nonvolatile EEPROM memory and has a maximal size 216
states. Thus, when a robot is switched off, it does not
lose its own information. The function of gene modiﬁcation
is accelerated by the hardware. In this way, genomes of
individual robots, through middleware system, build virtual
genome of the organism.
B. ENERGY HOMEOSTASIS OF A ROBOT ORGANISM
The preliminary experiments with symbiotic organisms
allowed us to detect a few critical issues for further develop-
ment. First of all, when robots aggregate into an organism,
they save energy. From this point of view, symbiotic life
form is more proﬁtable for robots. The targeted energy
homeostasis has the form similar to one, discussed in Sec. II-
B, only on the level of the whole organism. It is intended
the preprogrammed middleware takes care about internal
energy management, whereas the genome controls behavioral
strategies of ﬁnding energy sources. The organism should
develop these strategies evolutionary.
Swarm robots, when moving on the arena, have prepro-
grammed basic genes, such as ”move”, ”rotate”, ”dock in”,
”dock from”, ”actuate” and others. All basic genes and their
combinations have two weights: ”success” and ”potentially
consumed energy”. All these genes build a search space in
EEPROM. The gene sequences which leaded to successful
recharging are more weighted in ”success”. When some com-
bination does not lead to a success, its ”success” importance
is rapidly decreased. After each successful recharging, a
robot updates ”consumed energy”. It is also possible that
robots have a logbook of all updates for gene sequences.
Swarm robots ﬁrstly combine low-energy sequences, such
as ”move” or ”rotate”. When these low-energy sequences
have zero-near success weight (e.g. the barrier problem”),
robots look for further combinations of basic commands,
which have a higher energy consumption. In this way
several robots can aggregate together. In this aggregation
step we encountered the ﬁrst problem. All robots have
different genomes. Their simple combinations did not lead
to useful solutions [37], [40], [41]. Robots in the organism
start to develop their common virtual genome and only
then to introduce into this common genome their individual
information as a sequence of recessive genes. The problem
is even harder, when a robot already has a genome from
some past symbiotic forms. Currently it is not clear whether
all symbiotic robots converge to an unique genome with
different recessive sequences. One of the concept in this way
is to use the ”virtual robot sexuality” – robots, even during
disaggregated phase, exchange genetic information [38], see
Fig. 5.
Fig. 5.
Genome Mutation-Adding/overwriting via virtual robot
sexuality (from [37]).
Another encountered problem is a delayed ﬁtness response
of the robot organism. Typical life-cycle of swarm robots on
the 110×140cm arena is about a few minutes. The negative
feedback is issued, when during this time no energy source is
found. The life-cycle of the symbiotic form is much larger.
This means that a positive feedback can be achieved only
after a large number of unsuccessful steps, some of them
can even destroy the organism.
V. CONCLUSION
In this short paper we have shown a new paradigm in
collective systems, where the swarm robots get capable
of self-assembling into a single multi-robot organism. We
introduced an energy foraging scenario for both robot species
and demonstrated that a transition between collective and
symbiotic robot forms represents a hard problem. It involves
not only hardware and software issues, but also very basic
questions being also open in biological sense. We developed
a prototype of symbiotic robots capable for working in both,
swarm and organism, scenarios. Based on these works, fully
functional symbiotic robots will be developed [7].
REFERENCES
[1] D.W. Stephens and J.R. Krebs. Foraging Theory. Princeton University
Press, 1987.
[2] E. Bonabeau, M. Dorigo, and G. Theraulaz. Swarm intelligence: from
natural to artiﬁcial systems. Oxford University Press, New York, 1999.
[3] E. Sahin. Swarm Robotics: From sources of inspiration to domains of
application. Springer-Verlag, Heidelberg, Germany, 2004.
[4] Serge Kernbach, editor. Handbook of Collective Robotics: Fundamen-
tals and Challenges. Pan Stanford Publishing, Singapore, 2011.
[5] S. Kernbach, R. Thenius, O. Kernbach, and T. Schmickl.
Re-
embodiment of honeybee aggregation behavior in artiﬁcial micro-
robotic system. Adaptive Behavior, 17(3):237–259, 2009.
[6] H. Haken. Synergetics: An introduction, third edition. Springer-Verlag,
New York, 1983.
[7] Paul Levi and Serge Kernbach, editors.
Symbiotic Multi-Robot
Organisms: Reliability, Adaptability, Evolution. Springer Verlag, 2010.
[8] S. Kernbach, L. Ricotti, J. Liedke, P. Corradi, and M. Rothermel.
Study of macroscopic morphological features of symbiotic robotic
organisms.
In Proceedings of the workshop on self-reconﬁgurable
robots, IROS08, Nice, pages 18–25, 2008.
[9] S. Kornienko, O. Kornienko, and P. Levi.
Minimalistic approach
towards communication and perception in microrobotic swarms. In
Proc. of the International Conference on Intelligent Robots and
Systems (IROS-2005), pages 2228–2234, Edmonton, Canada, 2005.
[10] S. Kernbach, V. Nepomnyashchikh, T. Kancheva, and O. Kernbach.
Specialization and generalization of robotic behavior in swarm energy
foraging.
Mathematical and Computer Modelling of Dynamical
Systems, DOI 10.1080/13873954.2011.601421, 2011.
[11] S. Kernbach. Swarmrobot.org - Open-hardware Microrobotic Project
for Large-scale Artiﬁcial Swarms. ArXiv e-prints, arXiv1110.5762K,
October 2011.
[12] Serge Kernbach, Eugen Meister, Florian Schlachter, Kristof Jebens,
Marc Szymanski, Jens Liedke, Davide Laneri, Lutz Winkler, Thomas
Schmickl, Ronald Thenius, Paolo Corradi, and Leonardo Ricotti.
Symbiotic robot organisms: REPLICATOR and SYMBRION projects.
In Proceedings of the 8th Workshop on Performance Metrics for
Intelligent Systems, PerMIS ’08, pages 62–69, New York, NY, USA,
2008. ACM.
[13] L. K¨onig, K. Jebens, Serge Kernbach, and Paul Levi. Stability of on-
line and on-board evolving of adaptive collective behavior. In Herman
Bruyninckx, Libor Preucil, and Miroslav Kulich, editors, European
Robotics Symposium 2008, pages 293–302. 2008.
[14] M.C. Silverman, D. Nies, B. Jung, and G.S. Sukhatme. Staying alive:
A docking station for autonomous robot recharging. In Proc. of the
IEEE International Conference on Robotics and Automation, pages
1050–1055, Washington D.C., 2002.
[15] S. Kernbach and O. Kernbach. Collective energy homeostasis in a
large-scale micro-robotic swarm. Robotics and Autonomous Systems,
DOI 10.1016/j.robot.2011.08.001, 59:1090–1101, 2011.
[16] S. Camazine, J-L. Deneubourg, N.R. Franks, J. Sneyd, G. Theraulaz,
and E. Bonabeau. Self-Organization in Biological Systems. Princeton
University Press, Princeton, NJ, USA, 2003.
[17] S. Kornienko, O. Kornienko, and P. Levi. Swarm embodiment - a new
way for deriving emergent behaviour in artiﬁcial swarms. In P.Levi
and et al., editors, Autonome Mobile Systeme (AMS’05), pages 25–32,
2005.
[18] K. Jebens.
Development of a docking approach for autonomous
recharging system for micro-robot ”Jasmine”. Studient Thesis, Uni-
versity of Stuttgart, Germany, 2006.
[19] S. Kernbach. Structural Self-organization in Multi-Agents and Multi-
Robotic Systems. Logos Verlag, Berlin, 2008.
[20] Serge Kernbach, Heiko Hamann, J¨urgen Stradner, Ronald Thenius,
Thomas Schmickl, Karl Crailsheim, A. C. van Rossum, Michele
Sebag, Nicolas Bredeche, Yao Yao, Guy Baele, Yves Van de Peer,
Jon Timmis, Maizura Mohktar, Andy Tyrrell, A. E. Eiben, S. P.
McKibbin, Wenguo Liu, and Alan F. T. Winﬁeld. On adaptive self-
organization in artiﬁcial robot organisms.
In Proceedings of the
2009 Computation World: Future Computing, Service Computation,
Cognitive, Adaptive, Content, Patterns, COMPUTATIONWORLD ’09,
pages 33–43, Washington, DC, USA, 2009. IEEE Computer Society.
[21] S. Kornienko, O. Kornienko, and P. Levi. About nature of emergent
behavior in micro-systems. In Proc. of the Int. Conf. on Informatics in
Control, Automation and Robotics (ICINCO 2004), Setubal, Portugal,
pages 33–40, 2004.
[22] S. Kornienko, O. Kornienko, and P. Levi.
Flexible manufacturing
process planning based on the multi-agent technology. In Proc. of the
21st IASTED Int. Conf. on AI and Applications (AIA ’2003), Innsbruck,
Austria, pages 156–161, 2003.
[23] S. Kornienko, O. Kornienko, and P. Levi.
Multi-agent repairer of
damaged process plans in manufacturing environment. In Proc. of the
8th Conf. on Intelligent Autonomous Systems (IAS-8), Amsterdam, NL,
pages 485–494, 2004.
[24] S. Kornienko, O. Kornienko, C. Constantinescu, M. Pradier, and
P. Levi. Cognitive micro-agents: individual and collective perception
in microrobotic swarm. In Proc. of the IJCAI-05 Workshop on Agents
in real-time and dynamic environments, Edinburgh, UK, pages 33–42,
2005.
[25] S. Kornienko and O. Kornienko.
IR-based communication and
perception in microrobotic swarms. In IROS 2005, Edmonton, Canada,
2005.
[26] O. Kornienko, S. Kornienko, and P. Levi. Collective decision making
using natural self-organization in distributed systems.
In Proc. of
Int. Conf. on Computational Intelligence for Modelling, Control and
Automation (CIMCA’2001), Las Vegas, USA, pages 460–471, 2001.
[27] G. Weiss, editor.
Multiagent systems. A modern approach to dis-
tributed artiﬁcial intelligence. MIT Press, 1999.
[28] P. Levi, M. Schanz, S. Kornienko, and O. Kornienko. Application of
order parameter equation for the analysis and the control of nonlinear
time discrete dynamical systems.
Int. J. Bifurcation and Chaos,
9(8):1619–1634, 1999.
[29] D. H¨abe. Bio-inspired approach towards collective decision making
in robotic swarms. Master Thesis, University of Stuttgart, Germany,
2007.
[30] T. Kancheva. Adaptive role dynamics in energy foraging behavior of
a real micro-robotic swarm. Master Thesis, University of Stuttgart,
Germany, 2007.
[31] A. Attarzadeh.
Development of advanced power management for
autonomous micro-robots.
Master Thesis, University of Stuttgart,
Germany, 2006.
[32] A. Ishiguro and T. Maegawa.
Self-assembly through the interplay
between control and mechanical systems.
In Proc. of IEEE/RSJ06
Int. Conf. on Intelligent Robots and Systems, pages 631–638, Beijing,
China, 2006. IEEE.
[33] S. Murata, K. Kakomura, and H. Kurokawa. Docking experiments of a
modular robot by visual feedback. In Proc. of IEEE/RSJ06 Int. Conf.
on Intelligent Robots and Systems, pages 625–630, Beijing, China,
2006. IEEE.
[34] S. Kornienko, O. Kornienko, and P. Levi.
Generation of desired
emergent behavior in swarm of micro-robots. In R. Lopez de Mantaras
and L. Saitta, editors, Proc. of the 16th European conference on
artiﬁcial intelligence (ECAI 2004), Valencia, Spain, pages 239–243.
Amsterdam: IOS Press, 2004.
[35] SYMBRION. Symbiotic Evolutionary Robot Organisms, 7th Frame-
work Programme Project No FP7-ICT-2007.8.2. European Commu-
nities, 2008-2012.
[36] REPLICATOR.
Robotic Evolutionary Self-Programming and Self-
Assembling Organisms, 7th Framework Programme Project No FP7-
ICT-2007.2.1. European Communities, 2008-2012.
[37] A. Nagarathinam.
Development of a Software Framework for Jas-
mine Robots using ZigBee Communication Protocol. Master Thesis,
University of Stuttgart, Germany, 2007.
[38] Ch. Schwarzer. Investigation of Evolutionary Reproduction in a Robot
Swarm. Master Thesis, University of Stuttgart, Germany, 2008.
[39] Bruce Alberts, Alexander Johnson, Julian Lewis, Martin Raff, Keith
Roberts, and Peter Walter. Molecular Biology of the Cell. Garland
Science, March 2002.
[40] L. Koenig.
A model for developing behavioral patterns on multi-
robot organisms using concepts of natural evolution. Master Thesis,
University of Stuttgart, Germany, 2007.
[41] G. Speidel. Artiﬁcial cell differentiation in a multi-robot organisms
through gene regulation.
Master Thesis, University of Stuttgart,
Germany, 2008.