Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E.

April 19-21, 2005

UBIQUITOUS ROBOT: THE THIRD GENERATION OF ROBOTICS Jong-Hwan Kim, Kang-Hee Lee, and Yong-Duk Kim Robot Intelligence Technology Laboratory, Dept. of EECS, KAIST, Guseong-dong, Yuseong-gu, Daejeon, 305-701, Republic of Korea {johkim,khlee,ydkim}@rit.kaist.ac.kr
ABSTRACT This paper introduces ubiquitous robot (Ubibot) as a third generation of robotics, incorporating three forms of robots: software robot (Sobot), embeded robot (Embot) and mobile robot (Mobot), which can provide us with various services by any device through any network, at any place anytime in a ubiquitous space. Sobot is a virtual robot, which has the ability to move to any place through a network. Embot is embedded within the environment or in the Mobot. Mobot provides integrated mobile services, which are seamless, calm and context-aware. A Sobot, Rity, is introduced to investigate the usability of the proposed concept. Rity is a 3D synthetic character which exists in the virtual world, has a unique IP address and interacts with human beings through an Embot implemented by a face recognition system using a USB camera. To show the possibility of realization of Ubibot by using the current state of the art technologies, two kinds of successful demonstrations are presented. Also robot genome is proposed to implement a genetic robot, which is to investigate The Origin of Articial Species. To implement the robot genome, articial chromosome is introduced. This paper shows the personality of genetic robots is decided by their genome. 1. INTRODUCTION This paper is to investigate the feasibility of implementation of ubiquitous robot (Ubibot) by using the current state of the art technology, which can be dened as a third generation of robotics. Also it is to dene genetic robot to investigate The Origin of Articial Species. The genetic robot can be considered as an articial creature created by articial chromosome. In an ubiquitous era we will be living in a world where all objects such as electronic appliances are networked to each other and a robot will provide us with various services by any device through any network, at any place anytime. This robot is dened as a ubiquitous robot, Ubibot, which incorporates three forms of robots: software robot (Sobot), embeded robot (Embot) and mobile robot (Mobot) [1, 2, 3]. The Ubibot is following the paradigm shift of computer technology. The paradigm shift of robotics is motivated by ubiquitous computing and the evolution of computer technology in terms of the relationship between the technology and humans [4, 5]. Considering the evolution of robot technology, the rst generation was dominated by industrial robots followed by the second generation in which personal robots are becoming widespread these days, and as a third generation in the near future, Ubibot will appear. Comparing the paradigm change between the personal robot and ubiquitous robot eras, the former is based on individual robot systems and the latter will be employing multiple robot systems using real time broadband wireless network based on IPv6. The Ubibot has been developed based on the robot technology and the concept of ubiquitous computing in the Robot Intelligence Technology (RIT) Lab., KAIST since 2,000. In the future we will live in a ubiquitous world where all objects and devices are networked. In this ubiquitous space, u-space, a Ubibot will provide us with various services anytime, at any place, by any device, through any network. Following the general concepts of ubiquitous computing, Ubibot will be seamless, calm, context-aware, and networked. Although this paper is to investigate the feasibility of Ubibot, the basic concept leads us to seek the essence of what it means to be robot in the third generation of robotics. So far most of robot researchers have been devoted to develop and improve functionalities of robot such as intelligence, human-robot interaction, and mobility, without mentioning the essence of the robot as an articial creature. Since The Origin of Species by Charles Darwin in 1859, the concept of evolution has been widely spread around the world. Motivated by his discovery of evolution, the simulated evolution has been applied to engineering problems to get an optimal solution by employing the concept of chromosome representing the solution candidate. In 1976, Richard Dawkins claimed that We and other animals are machines created by our genes [6]. In the third generation of robotics, genetic robot can be proposed, which is created by articial chromosome [7]. This paper introduces a new concept of articial chromosome as the essence to dene the personality of a genetic robot and to pass on its traits to the next generation, like a genetic inheritance. It is an essential component for simulated evolution, which necessarily denes The Origin of Articial Species. If we think the origin in terms of the essence of the articial creatures, the essence should be a computerized genetic code, which determines a genetic robots propensity to feel happy, sad, angry, sleepy, hungry or afraid. The rst part of this paper presents the denition and basic concepts of Ubibot incorporating three forms of robots; Sobot, Embot, and Mobot. A Sobot, called Rity, developed at the RIT Lab., KAIST, is introduced to investigate the usability of the proposed concept of Ubibot [8, 9]. Rity is a 3D synthetic character which exists in the virtual world, has a unique IP address and interacts with human beings through an Embot implemented by a face recognition system using a USB camera. Rity is an autonomous agent which behaves based on its own internal states, and can interact with a person in real-time. It can provide us with an entertainment or a help through various interac-

AUS-ISM/05-1

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

tions in real life. To realize this, it needs an autonomous function, articial emotional model, learning skill, sociableness, and its own personality [10, 11]. It can be used as a character on a game or a movie or for the purpose of education [12, 13]. An architecture of Rity can be divided into ve modules: perception module, internal state module to implement motivation, homeostasis, and emotion [14, 15, 16], behavior selection module [17, 18], interactive learning module [19], and motor module. To show the possibility of realization of Ubibot, two kinds of demonstrations are carried out by using the current state of the art technologies. In the second part, Rity is considered as an articial creature living in a virtual world of a PC. It is a genetic robot which has its own genetic information. Rity is employed to test the worlds rst robotic chromosomes, which is a set of computerized DNA (Deoxyribonucleic acid) code for creating genetic robots that can think, feel, reason, express desire or intention, and could ultimately reproduce their kind, and evolve as a distinct species. Using this concept, a way to build articial chromosomes is proposed for genetic robots that would be capable of human-style evolution. Thus, the genetic code should be designed to represent all the traits and personality of articial creature: a manner of response to stimuli, the desire to avoid unpleasantness, to achieve intimacy and control, to satisfy curiosity or greed, and to prevent boredom, feelings of happiness, sadness, anger and fear to stimuli, and states of fatigue, hunger, drowsiness and so on, in order to imbue the articial creature with life. It can react emotionally to its environment, learns and makes reasoned decisions, based on an individual personality. The programmed genetic code is modelled on human DNA, though equivalent to a single strand of genetic code rather than the complex double helix of a real chromosome. The main functions of the genetic code are reproduction and evolution. This paper is organized as follows. Section II presents the denition and basic concepts of Ubibot. Section III describes the overall architecture of the Sobot. Demonstrations of the Sobot, Rity are provided in Section IV. Section V proposes genetic robot. Finally, concluding remarks follow in Section VI. 2. UBIQUITOUS ROBOT: UBIBOT Ubibot is a general term for all types of robots incorporating software robot (Sobot), embedded robot (Embot), and mobile robot (Mobot) which exist in a u-space. Ubibot exists in the u-space which provides physical and virtual environments. 2.1. U-space and Ubibot Ubiquitous space (u-space) is an environment in which ubiquitous computing is realized and every device is networked. The world will be composed of millions of u-spaces, each of which will be closely connected through ubiquitous networks. A robot working in a u-space is dened as a Ubibot and provides various services through any network by anyone at anytime and anywhere in a uspace. Ubibot in a u-space consists of both software and hardware robots. Sobot is a type of a software system whereas Embot and Mobot are hardware systems, Figure 1. Embots are located within the environment, human or otherwise, and are embedded in many devices. Their role is to sense, analyze and convey information to other Ubibots. Mobots are mobile robots. They can move

Figure 1: Ubibot in ubiquitous space

both independently and cooperatively, and provide practical services. Each Ubibot has specic individual intelligence and roles, and communicates information through networks. Sobot is capable of operating as an independent robot but it can also become the master system, which controls other Sobots, Embots and Mobots residing in other platforms as slave units. Their characteristics are summarized in the following. For details, the reader is referred to [2].

2.2. Software Robot: Sobot Since Sobot is software-based, it can easily move within the network and connect to other systems without any time or geographical limitation. It can be aware of situations and interact with the user seamlessly. Sobot can be introduced into the environment or other Mobots as a core system. It can control or, at an equal level, cooperate with Mobots. It can operate as an individual entity, without any help from other Ubibots. Sobot has three main characteristics, such as self-learning, context-aware intelligence, and calm and seamless interaction.

2.3. Embedded Robot: Embot EmBot is implanted in the environment or in Mobots. In cooperation with various sensors, Embot can detect the location of the user or a Mobot, authenticate them, integrate assorted sensor information and understand the environmental situation. An Embot may include all the objects which have both network and sensing functions, and be equipped with microprocessors. Embots generally have three major characteristics, such as calm sensing, information processing, and communication.

2.4. Mobile robot: Mobot Mobot is able to offer both a broad range of services for general users and specic functions within a specic u-space. Operating in u-space, Mobots have mobility as well as the capacity to provide general services in cooperation with Sobots and neighboring Embots. Mobot has the characteristics of manipulability by implementing arms and mobility which can be implemented in various types, such as wheel and biped. Mobot actions provide a broad range of services, such as personal, public, or eld services.

AUS-ISM/05-2

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

Sobot
Sensors
Vision Sound Tactile Gyro IR Timer

Perception module
Attention selector Symbolizer
Symbol vector Reward/penalty signal

Preference learner

Voice learner

Learning module

Motivation
Curiosity Sensor value Intimacy Monotony

Homeostasis
Fatigue Hunger Drowsiness

Emotion
Happiness Sadness Anger

3. IMPLEMENTATION OF SOBOT Sobot is a software robot which recognizes a situation by itself, behaves based on its own internal state, and can interact with a person in real-time. Sobot should be autonomous; it must be able to select a proper behavior according to its internal state such as motivation, homeostasis and emotion. Also, Sobot should be adaptable; it should adapt itself to its environment. For the purpose of achieving these functions easily and efciently, Sobot mimics an animal which is an autonomous and adaptable agent in nature. Fig. 2 shows an internal architecture of the proposed Sobot, Rity, where necessary modules are dened as follows: 1) Perception module, which perceives environment through virtual and physical sensors, 2) Internal state module, which includes motivation,homeostasis and emotion, 3) Behavior selection module, which selects a proper behavior, 4) Learning module, which learns from the interaction with a people, and 5) Motor module, which executes a behavior and expresses emotion.

3.1. Perception module The perception module includes a sensor unit, a releaser having stimulus information provided by a symbol vector and a sensitivity vector, and attention selector. This module can perceive and assess the environment and send the stimulus information to the internal state module. Sobot has several virtual sensors for light, sound, temperature, touch, vision, gyro, and time. Sobot can perceive 47 types of stimulus information from these sensors. Based on these information, Sobot can perform 77 different behaviors.

Virtual environment

Internal state module

Inherent behavior selector

Masking
Urgent flag

Behavior module
Behavior selector
Behavior End signal

Behavior

Actuator

Motor module

Figure 2: Internal architecture of Rity

3.2. Internal state module The internal state module denes the internal state with the motivation unit, the homeostasis unit and the emotion unit. Motivation (M ) is composed of six states: curiosity, intimacy, monotony, avoidance, greed, and the desire to control. Homeostasis (H) includes three states: fatigue, hunger, and drowsiness. Emotion (E) includes ve states: happiness, sadness, anger, fear, and neutral. According to the internal state, a proper behavior is selected.

3.3. Behavior selection module Behavior selection module is used to choose a proper behavior, based on Sobots internal state as well as stimulus information. When there is no command input from a user, various behaviors can be selected probabilistically by introducing a voting mechanism, where each behavior has its own voting value. The algorithm is described as follows: 1) Determine temporal voting vector, Vt using M and H, 2) Calculate voting vector V by masking Vt with attention command and emotion masks, 3) Calculate a behavior selection probability, p(b), using V , 4) Select a proper behavior b by p(b) among various behaviors. Initially, the temporal voting vector is calculated from the motivation and homeostasis as follows:

T Vt = MT DM + HT DH

(1)

=[vt1 , vt2 , , vtn ]

AUS-ISM/05-3

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

DM

dM 11 d = M 21 . . . dM x1 dH11

dM 12 dM 22

..

dM 1n . . . dH1n . . . dHyn dM xn (2)

dH12 dH22

..

d DH = H21 . . . dHy1

Even if a behavior is selected by both internal state and sensor information, there are still some limits on providing Sobot with natural behaviors. Inherent behavior selector makes up for the weak points in the behavior selector. It imitates an animals instinct. For instance, as soon as an obstacle like a wall or a cliff is found, it makes Sobot react to this situation immediately. Since it uses only sensory information directly, its decision making speed is faster than that of the behavior selector. The deterministic inherent behavior selector and the probabilistic behavior selector are complementary to each other for realizing a natural behavior. This means that it can help Sobot do right thing while carrying out various behaviors. 3.4. Motor module The motor module incorporates an actuator to execute behaviors and present emotions subject to the situation. 3.5. Learning module Learning module consists of preference learner and command learner. The former is to teach Sobot likes and dislikes for an object. If Sobot gets a reward or a penalty, the connected weights from the symbol to internal states are changed. On the other hand, the latter is to teach Sobot to do an appropriate behavior which a user wants Sobot to do. The learning can be considered as adjusting weighting parameters between commands and behaviors; if Sobot does a proper behavior for a given command, the weight between the command and the behavior is strengthened, and others are weakened. However, there are usually tens of behaviors. Thus, the learning process requires lot of time. Also it may be difcult to expect a desired behavior for an ordered command. To solve these problems, analogous behaviors are grouped into a subset before learning. For instance, the set SIT is composed of behaviors such as sit, crouch, and lie, and so on, as similar behaviors to sit. If a proper behavior is carried out for a certain command, all the corresponding weights of the subset are strengthened and vice versa. The update law is as follows: Wij (t + 1) = Wij (t) + Ri Ri = +Cr Cp on reward on penalty (6)

where n, x and y are the numbers of behaviors, motivations, and homeostases. vtk , k = 1, , n, is the temporal voting value, DM and DH are weights connecting the motivation and homeostasis to behaviors, respectively. As a next step, various maskings to the temporal voting vector, Vt are implemented considering emotion and external sensor information. Here, three kinds of masking are implemented to the temporal voting vector. These three kinds of maskings are masking for attention, masking for command, and masking for emotion. The masking process is to select a more appropriate behavior such that it prevents Sobot from carrying out unusual behaviors. For example, a behavior when it recognizes a ball should be different from that when it recognizes a person. When Sobot does not see the ball, masking for attention to the ball is carried out such that behaviors related to the ball are masked out and are not activated. An attention masking matrix Qa (Sa (t)) is obtained by the attention symbol, Sa (t). Each attention symbol has its own masking value and the matrix is dened as follows: Q (Sa (t)) =
a a q1 (Sa (t))

0
a q2 (Sa (t))

..

0 . . . 0

0 . . .
a qn (Sa (t))

(3) where n is a number of behaviors, q a () is a masking value, and 0 q a () 1. Similarly, command and emotion masking matrices are dened. From these three masking matrices and the temporal voting vector, the behavior selector obtains a nal voting vector as follows:
T VT =Vtemp Qa (a)Qv (c)Qe (e) =[v1 , v2 , , vn ]

(4)

where vk , k = 1, 2, , n, is the kth behaviors voting value. Finally, the selection probability p(bi ) of a behavior, bi , i = 1, 2, , n, is calculated from the voting values as follows: p(bi ) = vi
n k=1

(5)

(vk )

where Wij is a weight between the ith command and the jth behavior subset, is an emotion parameter, Ri is for a weight change for reward or penalty, and Cr and Cp are positive constants. When Sobot receives a patting (hitting) through a tactile sensor or a praise (a scolding) through a sound sensor, the perception module translates it as a reward (penalty). Weight is increased on reward, and decreased on penalty as shown in (6). It should be noted that an emotion parameter, is employed to consider the fact that learning rate depends on internal states. That is, learning speed is fast when happiness value is high and vice versa. Although the learning has been done on a behavior subset level, considering the direct contribution of the selected behavior the command masking values are assigned differently as follows:
v qm (ci ) = Wij q v (ci ) = Wij

By using the probability-based selection mechanism, the behavior selector can show diverse behaviors.

(7)

AUS-ISM/05-4

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005 with > > 0 v where qm (ci ) is a masking value of a behavior, bm carried out just now by the command, Ci and q v (ci ) indicates masking values of other behaviors in the same subset, Bi and and are positive constants. The command masking matrix is updated in proportion to weight values. A behavior activated just now and other behaviors in the same subset inuence different weight changes by and . Since is bigger than , the activated behavior gets a larger weight value than others in the same subset. 4. DEMONSTRATIONS FOR UBIBOT To demonstrate the usability of Rity for Ubibot, a Sobot, Rity is developed in a 3D virtual world. The following two demonstrations show seamless and omni-presence properties of Sobot. 4.1. Seamless integration of real and virtual worlds This section will demonstrate how, in a virtual environment, Rity will continuously cooperate with the real world with the help of a USB camera. The face recognition system stored in a PC watches the neighboring environment through the USB camera and, when a human is detected, analyzes, recognizes and authenticates the face. The result is to be sent to Rity through the network. Sobot will then react to the vision input information as it would normally react using the virtual sensing information. If the human is Ritys master, Rity will tend to stare at the master and happily greet him/her. Fig. 3, 4 and 5 are photographs of computer screens showing the virtual pet, Rity, in a virtual 3D environment. The small window at the bottom right of Fig. 3 shows the visual information in the form of a recognized face. A PCA method[20], which has been enhanced based on the evolutionary algorithm, was used for face detection. The window at the top right shows the graphical representation of the internal states of Rity.
Rity Ritys internal state

Figure 4: When Rity recognizes its master

Figure 5: When Rity detects a stranger

4.2. Omni-present Sobot This section discusses how Sobot can be connected and transmitted any time and at any place. Fig. 6 shows the interaction between Sobot A, owned by User A and Sobot B, owned by User B. For example, Sobot A is implemented at a local site, connects to the network and then invites Sobot B, located at a remote site, into its local space. Both Sobots (A and B) should have their own individual IP addresses. The User B will type in the ID and password and the IP address of Sobot B in order to access the remote site. Once access is approved, Sobot B, carrying its native characteristics and behavior patterns, can enter the local environment of User A. In Fig. 7, there are two Sobots in the local space. They look the same but have different characteristics. If the user gives the same stimulus to the two Sobots, for example, clicks once to pat or twice to hit, each Sobot will react differently because of their different characteristics. Fig. 7 shows the results of the experimentation after applying 10 instances of patting, or clicking, on both Sobots A and B. The gure shows the changes in internal states, facial expression and their behavior. As the amount of curiosity, intimacy and happiness increases, Sobot A starts moving around with a happy face, Fig. 7(a). On the other hand, in the case of Sobot B, the drowsiness increases making it sad and eventually sleepy. Fig. 7(b) and7(c) shows a comparison of the internal states of Sobot A and B.

Voice command Emotion Masters face recognition

Figure 3: Seamless integration of real and virtual worlds Fig. 4 shows an example, in which Rity recognizes its master. Rity then shows a happy look and welcomes him, with an increase of such internal states as curiosity, intimacy, and happiness. In Fig. 5, when a human who is not the master appears, Rity ignores him/her. In this case, for example, the internal state keeps as it has been.

AUS-ISM/05-5

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

(a)

(b)

Figure 6: Omni-present Sobot (a) connection with another Sobot in a remote site (b) IP address of a Sobot in a remote site, username and password for certication

(a)

Sobot can be downloaded and sent regardless of whether the site is local or remote. This is made possible by dening a common platform of the 3D graphic environment along with sensors and behaviors. 5. ROBOT GENOME This section presents a way to build an articial creature that would be capable of human-style evolution. As is well known, there are no one gene - one trait relationships in naturally evolved systems because of the pleiotypic and polygenic nature of the genotype, where pleiotypic nature has the effect that a single gene affects multiple phenotypic characters and polygenic nature has the effect that a single phenotypic character is affected by multiple genes. It means a single genetic change can affect every phenotypic characteristic. To reect this nature, unlike previously devised methods that associated stimuli with responses, Ritys chromosomal coding contains a sophisticated weighting system provided in the internal architecture (Figure 2). Internal relationships created by the weighting system allow Rity to be an individual capable of more than purely mechanistic response. Also it has a kind of programmed favoritism for one subtle shade of emotional or rational response, over another. Rity has fourteen articial chromosomes by which its traits can be passed on to its offspring. It perceives 47 different types of stimuli and can respond with 77 different behaviors as its responses. Figure 8 shows its 14 chromosomes, where chromosomes C1 -C6 are related to motivation, C7 -C9 to homeostasis, and C10 C14 to emotion. Also, motivation is composed of six states (chromosomes): curiosity (C1 ), intimacy (C2 ), monotony (C3 ), avoidance (C4 ), greed (C5 ), and desire to control (C6 ). Homeostasis includes three states (chromosomes): fatigue (C7 ), drowsiness (C8 ), and hunger (C9 ). Emotion includes ve states (chromosomes): happiness (C10 ), sadness (C11 ), anger (C12 ), fear (C13 ), and neutral (C14 ). Each chromosome Ck , k = 1, 2, , 14, consists of three kinds of genes: F-genes xF k , I-genes xI k , and B-genes xR k , dened as xF k I Ck = x k xB k
(b) (c)

Figure 7: Omni-presence (a) Sobot A in a local site and Sobot B downloaded from a remote site (b) Internal state of Sobot A (c) Internal state of Sobot B with I x1 xF 1 xI xF 2 2 I = . ,x k = . . . . . F xI xp q B x1 xB B 2 ,x k = . . . B xn

xF k

(9)

where p, q, and n are total numbers of F-genes, I-genes, and B-genes in a chromosome. In Rity, p = 5, q = 47, and n = 77. Consequently, a robot genome G, which means a chromosomal set with genetic codes determining Ritys personality, is dened as G= C1 C2 C14 . (10)

(8)

F-genes represents fundamental characteristics of Rity, including genetic information such as volatility, initial and mean value of each internal state, some intrinsic parameters, etc. Volatility means whether the internal state is volatile or non-volatile since operating point in time. F-genes can also include sex, life span, color and so on to dene its fundamental nature. I-genes includes genetic codes representing its internal preference by setting the weights between the stimulus and internal state. Theses genes shape variety of the internal state affected by stimuli and have information of whether the stimulus satises, amplies, or has nothing to do with the internal state. After the birth, the preference can be trained on-line by adapting the weights like pet training.

AUS-ISM/05-6

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

C1
F-GENES

C2

C3

C4

C5

C6

C7

C8

C9 C10 C11 C12 C13 C14

I-GENES

(a)
B-GENES

(b)

Figure 9: Two different chromosome set of Ritys: (a) chromosome set of Rity A and (b) chromosome set of Rity B

Figure 8: Articial chromosomes of Rity

B-genes includes genetic codes related to output behavior by setting the weights between the internal state and voting vector. These genes are in charge of behavior selection, its frequency, and its activation level based on the internal state. They also include masking information which prevents Rity from doing unnecessary emotional expression and behaviors. Genetic robot can be dened as a robot which has its own genetic codes. This section veries the concept of genetic robot by implanting the articial chromosomes into Ritys. The genes in Figure 8 are originally represented by real numbers; values of Fgenes range from 1 to 500, I-genes from 0 to 5000, and B-genes from 1 to 1000. Like a DNA analysis, these genes are normalized to brightness values from 0 to 255, which are expressed to blackand-white rectangles. The darker the color is, the higher its value is. Figure 9 shows two different chromosome set of Ritys. As per their genetic codes, no two Ritys react the same way to their surroundings as shown in Figure 10. One was bored; the other was panted and expressed happiness at the sight of their human handlers because they had a different personality. It totally depends on their genes. Current version is equivalent to a single strand of genetic code of real numbers rather than the complex double helix of a real chromosome. One of future works is on the equivalent of X and Y chromosomes that would confer sexual characteristics. Thus, if male and female like each other, they could have their own children. Also the software chromosomes will be implanted in a mobile robot so that they will imbue the robot with life. Considering the concept of ubiquitous robot and ubiquitous computing environment, however, a key future work should be that the robot genome is to be sent via the Internet to other computers or pieces of hardware, becoming a sort of wirelessly transmissible soul that would invisibly control the actions and desires of future interconnected appliances, from devices in a smart home or ofce, to cellphones or security cameras including ubibots.

(a)

(b)

(c)

Figure 10: Rity A and Rity B in a virtual space: (a) Rity A and Rity B, (b) Internal state of Rity A, and (c) Internal state of Rity B

AUS-ISM/05-7

Proceeding of the Second American University of Sharjah International Symposium on Mechatronics, Sharjah, U.A.E. April 19-21, 2005

6. CONCLUDING REMARKS This paper introduced a ubiquitous robot, Ubibot, as a third generation of robotics, which integrates three forms of robots: Sobot, Embot and Mobot. Rity, a Sobot or an articial creature living in a virtual 3D world of a PC, was implemented using two scenarios to demonstrate the possibility of realizing Ubibot. The rst scenario illustrated how Rity, with the support of Embot, could recognize its master and reacted properly. This was to show the seamless integration of real and virtual worlds. The second scenario demonstrated how Sobots could be transmitted through networks and be transposed into different locations. This was to demonstrate the omni-presence capability by using Sobot. This paper also proposed a new concept of robot genome to investigate The Origin of Articial Species, such as genetic robot. The robot genome was implanted into Rity to test a feasibility that genetic robots could have their own personality. The articial chromosomes will lead the genetic robot to reproduction and evolution. In the new ubiquitous era, our future world will be composed of millions of u-spaces, each of which will be closely connected through ubiquitous networks. In this u-space we can expect that Ubibot will help us whenever we click as Genie of the Magic Lamp did for Aladdin.

[8]

Y.-D. Kim, Y.-J. Kim, J.-H. Kim and J.-R. Lim, Implementaton of Articial Creature based on Interactive Learning, in Proc.of FIRA Robot World Congress, Seoul, Korea, pp. 369-373, May 2002. Y.-D. Kim, Y.-J. Kim, and J.-H. Kim, Behavior Selection and Learning for Synthetic Character, in Proceedings of the IEEE Congress on Evolutionary Computation, pp. 898-903, 2004.

[9]

[10] C. Breazeal, Function Meets Style: Insights From Emotion Theory Applied to HRI, IEEE Trans. on Systems, Man, and Cybernetics, Part C, vol.32, no. 2, pp. 187-194, May 2004. [11] H. Miwa, T. Umetsu, A. Takanishi, and H. Takanobu, Robot personality based on the equation of emotion dened in the 3d mental space, in Proc. of IEEE Int. Conf. on Robotics and Automation, vol. 3, Seoul, Korea, pp. 2602-2607, May 2001. [12] J. Bates, A.B. Loyall and W.S. Reilly, Integrating Reactivity, Goals, and Emotion in a Broad Agent, in Proc. of 14th Ann. Conf. Cognitive Science Soc., Bloomington, IN, July 1992. [13] M. Mateas, An Oz-Centric Review of Interactive Drama and Believable Agents, AI Today: Recent Trends and Developments, Lecture Notes in Articial Intelligence no. 1600, pp. 297-328, Springer-Verlag, Berlin, 1999. [14] C. Kline and B. Blumberg, The Art and Science of Synthetic Character Design, in Proc. of the AISB 1999 Symp. on AI and Creativity in Entertainment and visual Art, Edinburgh, Scotland, 1999. a [15] J.-D Vel squez, An emotion-based approach to robotics, in Proc. of IEEE/RSJ Int. Conf. on Intellighent Robots and Systems, vol. 1, Kyongju, Korea, Oct. 1999, pp. 235-240. [16] N. Kubota, Y. Nojima, N. Baba, F. Kojima, and T. Fukuda, Evolving pet robot with emotional model, in Proc. of IEEE Congress on Evolutionary Computation, vol. 2, San Diego, CA, pp. 1231-1237, July 2000. [17] R. C. Arkin, M. Fujita, T. Takagi and R. Hasehawa, Ethological Modeling and Architecture for an Entertainment Robot, in Proc. of IEEE Int. Conf. on Robotics and Automation, vol.1, Seoul, Korea, May 2001, pp. 453-458. [18] D. Isla, R. Burke, M. Downie, and B. Blumberg, A Layered Brain Architecture for Synthetic Creatures, in Proc. of the Int. Joint Conf. on Artical Intelligence, Seattle, WA, Aug. 2001, pp. 1051-1058. [19] S.-Y. Yoon, B. M. Blumberg, and G. E. Schneider, Motivation driven learning for interactive synthetic characters, in Proc. of the fourth Int. Conf. on Autonomous Agents, Barcelona, Spain, Jun. 2000, pp. 365-372. [20] J.-S. Jang, K.-H. Han, and J.-H. Kim, Face Detection using Quantum-inspired Evolutionary Algorithm, in Proc. of the Congress on Evolutionary Computation, Portland, OR, Jun. 2004, pp. 2100-2107.

Acknowledgments
This work was supported by the Ministry of information & Communications, Korea, under the Information Technology Research Center (ITRC) Support Program. 7. REFERENCES [1] Jong-Hwan Kim, IT-based UbiBot, in the Issue of the 13th of May, 2003, The Korea Electronic Times, Special Theme Lecture Article, Seoul, Korea, May 2003. Jong-Hwan Kim, Ubiquitous Robot, in Proc.of Fuzzy Days International Conference, Dortmund, Germany, September 2004, (Keynote Speech Paper). J.-H. Kim, Y.-D. Kim, and K.-H. Lee,The Third Generation of Robotics: Ubiquitous Robot,in Proc. of the International Conference on Autonomous Robots and Agents, Palmerston North, New Zealand, December 13, 2004 (Keynote Speech Paper) Mark Weiser, The computer for the 21st century, Scientic American, Vol. 265, No. 3, pp. 94-104, Sept. 1991. Mark Weiser, Some computer science problems in ubiquitous computing, Communications of ACM, Vol. 36, No.7, pp. 75-84, July 1993. R. Dawkins, The Selsh Gene, The Oxford Publication Press, 1976. J.-H. Kim, K.-H. Lee, Y.-D. Kim, B.-J. Lee, J.-K. Yoo and S.-H. Choi, The Origin of Articial Species: Humanoid Robot HanSaRam, in Proc. of the 2nd International Conference on Humanoid, Nanotechnology, Information technology, Communication and control, Environment, and Management (HNICEM05), Manila, Philippines, March 2005 (Plenary Speech Paper),

[2]

[3]

[4] [5]

[6] [7]

AUS-ISM/05-8